Merge branch 'master' of https://github.com/ClickHouse/ClickHouse into testflows_modules_in_parallel

base/common/DecomposedFloat.h

@@ -0,0 +1,216 @@
+#pragma once
+
+#include <cstdint>
+#include <cstddef>
+#include <cstring>
+#include <common/extended_types.h>
+
+
+/// Allows to check the internals of IEEE-754 floating point number.
+
+template <typename T> struct FloatTraits;
+
+template <>
+struct FloatTraits<float>
+{
+    using UInt = uint32_t;
+    static constexpr size_t bits = 32;
+    static constexpr size_t exponent_bits = 8;
+    static constexpr size_t mantissa_bits = bits - exponent_bits - 1;
+};
+
+template <>
+struct FloatTraits<double>
+{
+    using UInt = uint64_t;
+    static constexpr size_t bits = 64;
+    static constexpr size_t exponent_bits = 11;
+    static constexpr size_t mantissa_bits = bits - exponent_bits - 1;
+};
+
+
+/// x = sign * (2 ^ normalized_exponent) * (1 + mantissa * 2 ^ -mantissa_bits)
+/// x = sign * (2 ^ normalized_exponent + mantissa * 2 ^ (normalized_exponent - mantissa_bits))
+template <typename T>
+struct DecomposedFloat
+{
+    using Traits = FloatTraits<T>;
+
+    DecomposedFloat(T x)
+    {
+        memcpy(&x_uint, &x, sizeof(x));
+    }
+
+    typename Traits::UInt x_uint;
+
+    bool is_negative() const
+    {
+        return x_uint >> (Traits::bits - 1);
+    }
+
+    /// Returns 0 for both +0. and -0.
+    int sign() const
+    {
+        return (exponent() == 0 && mantissa() == 0)
+            ? 0
+            : (is_negative()
+                ? -1
+                : 1);
+    }
+
+    uint16_t exponent() const
+    {
+        return (x_uint >> (Traits::mantissa_bits)) & (((1ull << (Traits::exponent_bits + 1)) - 1) >> 1);
+    }
+
+    int16_t normalized_exponent() const
+    {
+        return int16_t(exponent()) - ((1ull << (Traits::exponent_bits - 1)) - 1);
+    }
+
+    uint64_t mantissa() const
+    {
+        return x_uint & ((1ull << Traits::mantissa_bits) - 1);
+    }
+
+    int64_t mantissa_with_sign() const
+    {
+        return is_negative() ? -mantissa() : mantissa();
+    }
+
+    /// NOTE Probably floating point instructions can be better.
+    bool is_integer_in_representable_range() const
+    {
+        return x_uint == 0
+            || (normalized_exponent() >= 0  /// The number is not less than one
+                /// The number is inside the range where every integer has exact representation in float
+                && normalized_exponent() <= static_cast<int16_t>(Traits::mantissa_bits)
+                /// After multiplying by 2^exp, the fractional part becomes zero, means the number is integer
+                && ((mantissa() & ((1ULL << (Traits::mantissa_bits - normalized_exponent())) - 1)) == 0));
+    }
+
+
+    /// Compare float with integer of arbitrary width (both signed and unsigned are supported). Assuming two's complement arithmetic.
+    /// Infinities are compared correctly. NaNs are treat similarly to infinities, so they can be less than all numbers.
+    /// (note that we need total order)
+    template <typename Int>
+    int compare(Int rhs)
+    {
+        if (rhs == 0)
+            return sign();
+
+        /// Different signs
+        if (is_negative() && rhs > 0)
+            return -1;
+        if (!is_negative() && rhs < 0)
+            return 1;
+
+        /// Fractional number with magnitude less than one
+        if (normalized_exponent() < 0)
+        {
+            if (!is_negative())
+                return rhs > 0 ? -1 : 1;
+            else
+                return rhs >= 0 ? -1 : 1;
+        }
+
+        /// The case of the most negative integer
+        if constexpr (is_signed_v<Int>)
+        {
+            if (rhs == std::numeric_limits<Int>::lowest())
+            {
+                assert(is_negative());
+
+                if (normalized_exponent() < static_cast<int16_t>(8 * sizeof(Int) - is_signed_v<Int>))
+                    return 1;
+                if (normalized_exponent() > static_cast<int16_t>(8 * sizeof(Int) - is_signed_v<Int>))
+                    return -1;
+
+                if (mantissa() == 0)
+                    return 0;
+                else
+                    return -1;
+            }
+        }
+
+        /// Too large number: abs(float) > abs(rhs). Also the case with infinities and NaN.
+        if (normalized_exponent() >= static_cast<int16_t>(8 * sizeof(Int) - is_signed_v<Int>))
+            return is_negative() ? -1 : 1;
+
+        using UInt = make_unsigned_t<Int>;
+        UInt uint_rhs = rhs < 0 ? -rhs : rhs;
+
+        /// Smaller octave: abs(rhs) < abs(float)
+        if (uint_rhs < (static_cast<UInt>(1) << normalized_exponent()))
+            return is_negative() ? -1 : 1;
+
+        /// Larger octave: abs(rhs) > abs(float)
+        if (normalized_exponent() + 1 < static_cast<int16_t>(8 * sizeof(Int) - is_signed_v<Int>)
+            && uint_rhs >= (static_cast<UInt>(1) << (normalized_exponent() + 1)))
+            return is_negative() ? 1 : -1;
+
+        /// The same octave
+        /// uint_rhs == 2 ^ normalized_exponent + mantissa * 2 ^ (normalized_exponent - mantissa_bits)
+
+        bool large_and_always_integer = normalized_exponent() >= static_cast<int16_t>(Traits::mantissa_bits);
+
+        typename Traits::UInt a = large_and_always_integer
+            ? mantissa() << (normalized_exponent() - Traits::mantissa_bits)
+            : mantissa() >> (Traits::mantissa_bits - normalized_exponent());
+
+        typename Traits::UInt b = uint_rhs - (static_cast<UInt>(1) << normalized_exponent());
+
+        if (a < b)
+            return is_negative() ? 1 : -1;
+        if (a > b)
+            return is_negative() ? -1 : 1;
+
+        /// Float has no fractional part means that the numbers are equal.
+        if (large_and_always_integer || (mantissa() & ((1ULL << (Traits::mantissa_bits - normalized_exponent())) - 1)) == 0)
+            return 0;
+        else
+            /// Float has fractional part means its abs value is larger.
+            return is_negative() ? -1 : 1;
+    }
+
+
+    template <typename Int>
+    bool equals(Int rhs)
+    {
+        return compare(rhs) == 0;
+    }
+
+    template <typename Int>
+    bool notEquals(Int rhs)
+    {
+        return compare(rhs) != 0;
+    }
+
+    template <typename Int>
+    bool less(Int rhs)
+    {
+        return compare(rhs) < 0;
+    }
+
+    template <typename Int>
+    bool greater(Int rhs)
+    {
+        return compare(rhs) > 0;
+    }
+
+    template <typename Int>
+    bool lessOrEquals(Int rhs)
+    {
+        return compare(rhs) <= 0;
+    }
+
+    template <typename Int>
+    bool greaterOrEquals(Int rhs)
+    {
+        return compare(rhs) >= 0;
+    }
+};
+
+
+using DecomposedFloat64 = DecomposedFloat<double>;
+using DecomposedFloat32 = DecomposedFloat<float>;

base/common/arithmeticOverflow.h

@@ -56,27 +56,33 @@ namespace common
     }
 
     template <>
-    inline bool addOverflow(__int128 x, __int128 y, __int128 & res)
+    inline bool addOverflow(Int128 x, Int128 y, Int128 & res)
     {
-        static constexpr __int128 min_int128 = minInt128();
-        static constexpr __int128 max_int128 = maxInt128();
         res = addIgnoreOverflow(x, y);
-        return (y > 0 && x > max_int128 - y) || (y < 0 && x < min_int128 - y);
+        return (y > 0 && x > std::numeric_limits<Int128>::max() - y) ||
+            (y < 0 && x < std::numeric_limits<Int128>::min() - y);
     }
 
     template <>
-    inline bool addOverflow(wInt256 x, wInt256 y, wInt256 & res)
+    inline bool addOverflow(UInt128 x, UInt128 y, UInt128 & res)
     {
         res = addIgnoreOverflow(x, y);
-        return (y > 0 && x > std::numeric_limits<wInt256>::max() - y) ||
-            (y < 0 && x < std::numeric_limits<wInt256>::min() - y);
+        return x > std::numeric_limits<UInt128>::max() - y;
     }
 
     template <>
-    inline bool addOverflow(wUInt256 x, wUInt256 y, wUInt256 & res)
+    inline bool addOverflow(Int256 x, Int256 y, Int256 & res)
     {
         res = addIgnoreOverflow(x, y);
-        return x > std::numeric_limits<wUInt256>::max() - y;
+        return (y > 0 && x > std::numeric_limits<Int256>::max() - y) ||
+            (y < 0 && x < std::numeric_limits<Int256>::min() - y);
+    }
+
+    template <>
+    inline bool addOverflow(UInt256 x, UInt256 y, UInt256 & res)
+    {
+        res = addIgnoreOverflow(x, y);
+        return x > std::numeric_limits<UInt256>::max() - y;
     }
 
     template <typename T>
@@ -104,24 +110,30 @@ namespace common
     }
 
     template <>
-    inline bool subOverflow(__int128 x, __int128 y, __int128 & res)
+    inline bool subOverflow(Int128 x, Int128 y, Int128 & res)
     {
-        static constexpr __int128 min_int128 = minInt128();
-        static constexpr __int128 max_int128 = maxInt128();
         res = subIgnoreOverflow(x, y);
-        return (y < 0 && x > max_int128 + y) || (y > 0 && x < min_int128 + y);
+        return (y < 0 && x > std::numeric_limits<Int128>::max() + y) ||
+            (y > 0 && x < std::numeric_limits<Int128>::min() + y);
     }
 
     template <>
-    inline bool subOverflow(wInt256 x, wInt256 y, wInt256 & res)
+    inline bool subOverflow(UInt128 x, UInt128 y, UInt128 & res)
     {
         res = subIgnoreOverflow(x, y);
-        return (y < 0 && x > std::numeric_limits<wInt256>::max() + y) ||
-            (y > 0 && x < std::numeric_limits<wInt256>::min() + y);
+        return x < y;
     }
 
     template <>
-    inline bool subOverflow(wUInt256 x, wUInt256 y, wUInt256 & res)
+    inline bool subOverflow(Int256 x, Int256 y, Int256 & res)
+    {
+        res = subIgnoreOverflow(x, y);
+        return (y < 0 && x > std::numeric_limits<Int256>::max() + y) ||
+            (y > 0 && x < std::numeric_limits<Int256>::min() + y);
+    }
+
+    template <>
+    inline bool subOverflow(UInt256 x, UInt256 y, UInt256 & res)
     {
         res = subIgnoreOverflow(x, y);
         return x < y;
@@ -151,36 +163,33 @@ namespace common
         return __builtin_smulll_overflow(x, y, &res);
     }
 
+    /// Overflow check is not implemented for big integers.
+
     template <>
-    inline bool mulOverflow(__int128 x, __int128 y, __int128 & res)
+    inline bool mulOverflow(Int128 x, Int128 y, Int128 & res)
     {
         res = mulIgnoreOverflow(x, y);
-        if (!x || !y)
-            return false;
-
-        unsigned __int128 a = (x > 0) ? x : -x;
-        unsigned __int128 b = (y > 0) ? y : -y;
-        return mulIgnoreOverflow(a, b) / b != a;
+        return false;
     }
 
     template <>
-    inline bool mulOverflow(wInt256 x, wInt256 y, wInt256 & res)
+    inline bool mulOverflow(Int256 x, Int256 y, Int256 & res)
     {
         res = mulIgnoreOverflow(x, y);
-        if (!x || !y)
-            return false;
-
-        wInt256 a = (x > 0) ? x : -x;
-        wInt256 b = (y > 0) ? y : -y;
-        return mulIgnoreOverflow(a, b) / b != a;
+        return false;
     }
 
     template <>
-    inline bool mulOverflow(wUInt256 x, wUInt256 y, wUInt256 & res)
+    inline bool mulOverflow(UInt128 x, UInt128 y, UInt128 & res)
     {
         res = mulIgnoreOverflow(x, y);
-        if (!x || !y)
-            return false;
-        return res / y != x;
+        return false;
+    }
+
+    template <>
+    inline bool mulOverflow(UInt256 x, UInt256 y, UInt256 & res)
+    {
+        res = mulIgnoreOverflow(x, y);
+        return false;
     }
 }

base/common/extended_types.h

@@ -5,16 +5,14 @@
 #include <common/types.h>
 #include <common/wide_integer.h>
 
-using Int128 = __int128;
 
-using wInt256 = wide::integer<256, signed>;
-using wUInt256 = wide::integer<256, unsigned>;
+using Int128 = wide::integer<128, signed>;
+using UInt128 = wide::integer<128, unsigned>;
+using Int256 = wide::integer<256, signed>;
+using UInt256 = wide::integer<256, unsigned>;
 
-static_assert(sizeof(wInt256) == 32);
-static_assert(sizeof(wUInt256) == 32);
-
-static constexpr __int128 minInt128() { return static_cast<unsigned __int128>(1) << 127; }
-static constexpr __int128 maxInt128() { return (static_cast<unsigned __int128>(1) << 127) - 1; }
+static_assert(sizeof(Int256) == 32);
+static_assert(sizeof(UInt256) == 32);
 
 /// The standard library type traits, such as std::is_arithmetic, with one exception
 /// (std::common_type), are "set in stone". Attempting to specialize them causes undefined behavior.
@@ -26,7 +24,7 @@ struct is_signed
 };
 
 template <> struct is_signed<Int128> { static constexpr bool value = true; };
-template <> struct is_signed<wInt256> { static constexpr bool value = true; };
+template <> struct is_signed<Int256> { static constexpr bool value = true; };
 
 template <typename T>
 inline constexpr bool is_signed_v = is_signed<T>::value;
@@ -37,7 +35,8 @@ struct is_unsigned
     static constexpr bool value = std::is_unsigned_v<T>;
 };
 
-template <> struct is_unsigned<wUInt256> { static constexpr bool value = true; };
+template <> struct is_unsigned<UInt128> { static constexpr bool value = true; };
+template <> struct is_unsigned<UInt256> { static constexpr bool value = true; };
 
 template <typename T>
 inline constexpr bool is_unsigned_v = is_unsigned<T>::value;
@@ -51,8 +50,9 @@ struct is_integer
 };
 
 template <> struct is_integer<Int128> { static constexpr bool value = true; };
-template <> struct is_integer<wInt256> { static constexpr bool value = true; };
-template <> struct is_integer<wUInt256> { static constexpr bool value = true; };
+template <> struct is_integer<UInt128> { static constexpr bool value = true; };
+template <> struct is_integer<Int256> { static constexpr bool value = true; };
+template <> struct is_integer<UInt256> { static constexpr bool value = true; };
 
 template <typename T>
 inline constexpr bool is_integer_v = is_integer<T>::value;
@@ -64,7 +64,11 @@ struct is_arithmetic
     static constexpr bool value = std::is_arithmetic_v<T>;
 };
 
-template <> struct is_arithmetic<__int128> { static constexpr bool value = true; };
+template <> struct is_arithmetic<Int128> { static constexpr bool value = true; };
+template <> struct is_arithmetic<UInt128> { static constexpr bool value = true; };
+template <> struct is_arithmetic<Int256> { static constexpr bool value = true; };
+template <> struct is_arithmetic<UInt256> { static constexpr bool value = true; };
+
 
 template <typename T>
 inline constexpr bool is_arithmetic_v = is_arithmetic<T>::value;
@@ -75,9 +79,10 @@ struct make_unsigned
     typedef std::make_unsigned_t<T> type;
 };
 
-template <> struct make_unsigned<Int128> { using type = unsigned __int128; };
-template <> struct make_unsigned<wInt256>  { using type = wUInt256; };
-template <> struct make_unsigned<wUInt256> { using type = wUInt256; };
+template <> struct make_unsigned<Int128> { using type = UInt128; };
+template <> struct make_unsigned<UInt128> { using type = UInt128; };
+template <> struct make_unsigned<Int256>  { using type = UInt256; };
+template <> struct make_unsigned<UInt256> { using type = UInt256; };
 
 template <typename T> using make_unsigned_t = typename make_unsigned<T>::type;
 
@@ -87,8 +92,10 @@ struct make_signed
     typedef std::make_signed_t<T> type;
 };
 
-template <> struct make_signed<wInt256>  { using type = wInt256; };
-template <> struct make_signed<wUInt256> { using type = wInt256; };
+template <> struct make_signed<Int128>  { using type = Int128; };
+template <> struct make_signed<UInt128> { using type = Int128; };
+template <> struct make_signed<Int256>  { using type = Int256; };
+template <> struct make_signed<UInt256> { using type = Int256; };
 
 template <typename T> using make_signed_t = typename make_signed<T>::type;
 
@@ -98,8 +105,10 @@ struct is_big_int
     static constexpr bool value = false;
 };
 
-template <> struct is_big_int<wInt256> { static constexpr bool value = true; };
-template <> struct is_big_int<wUInt256> { static constexpr bool value = true; };
+template <> struct is_big_int<Int128> { static constexpr bool value = true; };
+template <> struct is_big_int<UInt128> { static constexpr bool value = true; };
+template <> struct is_big_int<Int256> { static constexpr bool value = true; };
+template <> struct is_big_int<UInt256> { static constexpr bool value = true; };
 
 template <typename T>
 inline constexpr bool is_big_int_v = is_big_int<T>::value;

base/common/itoa.h

@@ -30,9 +30,8 @@
 #include <cstddef>
 #include <cstring>
 #include <type_traits>
+#include <common/extended_types.h>
 
-using int128_t = __int128;
-using uint128_t = unsigned __int128;
 
 namespace impl
 {
@@ -106,7 +105,7 @@ using UnsignedOfSize = typename SelectType
     uint16_t,
     uint32_t,
     uint64_t,
-    uint128_t
+    __uint128_t
 >::Result;
 
 /// Holds the result of dividing an unsigned N-byte variable by 10^N resulting in
@@ -313,7 +312,8 @@ namespace convert
     }
 }
 
-static inline int digits10(uint128_t x)
+template <typename T>
+static inline int digits10(T x)
 {
     if (x < 10ULL)
         return 1;
@@ -346,8 +346,11 @@ static inline int digits10(uint128_t x)
     return 12 + digits10(x / 1000000000000ULL);
 }
 
-static inline char * writeUIntText(uint128_t x, char * p)
+template <typename T>
+static inline char * writeUIntText(T x, char * p)
 {
+    static_assert(is_unsigned_v<T>);
+
     int len = digits10(x);
     auto pp = p + len;
     while (x >= 100)
@@ -370,14 +373,28 @@ static inline char * writeLeadingMinus(char * pos)
     return pos + 1;
 }
 
-static inline char * writeSIntText(int128_t x, char * pos)
+template <typename T>
+static inline char * writeSIntText(T x, char * pos)
 {
-    static constexpr int128_t min_int128 = uint128_t(1) << 127;
+    static_assert(std::is_same_v<T, Int128> || std::is_same_v<T, Int256>);
 
-    if (unlikely(x == min_int128))
+    using UnsignedT = make_unsigned_t<T>;
+    static constexpr T min_int = UnsignedT(1) << (sizeof(T) * 8 - 1);
+
+    if (unlikely(x == min_int))
     {
-        memcpy(pos, "-170141183460469231731687303715884105728", 40);
-        return pos + 40;
+        if constexpr (std::is_same_v<T, Int128>)
+        {
+            const char * res = "-170141183460469231731687303715884105728";
+            memcpy(pos, res, strlen(res));
+            return pos + strlen(res);
+        }
+        else if constexpr (std::is_same_v<T, Int256>)
+        {
+            const char * res = "-57896044618658097711785492504343953926634992332820282019728792003956564819968";
+            memcpy(pos, res, strlen(res));
+            return pos + strlen(res);
+        }
     }
 
     if (x < 0)
@@ -385,7 +402,7 @@ static inline char * writeSIntText(int128_t x, char * pos)
         x = -x;
         pos = writeLeadingMinus(pos);
     }
-    return writeUIntText(static_cast<uint128_t>(x), pos);
+    return writeUIntText(UnsignedT(x), pos);
 }
 
 }
@@ -403,13 +420,25 @@ inline char * itoa(char8_t i, char * p)
 }
 
 template <>
-inline char * itoa<uint128_t>(uint128_t i, char * p)
+inline char * itoa(UInt128 i, char * p)
 {
     return impl::writeUIntText(i, p);
 }
 
 template <>
-inline char * itoa<int128_t>(int128_t i, char * p)
+inline char * itoa(Int128 i, char * p)
+{
+    return impl::writeSIntText(i, p);
+}
+
+template <>
+inline char * itoa(UInt256 i, char * p)
+{
+    return impl::writeUIntText(i, p);
+}
+
+template <>
+inline char * itoa(Int256 i, char * p)
 {
     return impl::writeSIntText(i, p);
 }

base/common/strong_typedef.h

@@ -4,7 +4,8 @@
 #include <type_traits>
 #include <utility>
 
-template <class T, class Tag>
+
+template <typename T, typename Tag>
 struct StrongTypedef
 {
 private:
@@ -38,14 +39,16 @@ public:
 
     bool operator==(const Self & rhs) const { return t == rhs.t; }
     bool operator<(const Self & rhs) const { return t < rhs.t; }
+    bool operator>(const Self & rhs) const { return t > rhs.t; }
 
     T & toUnderType() { return t; }
     const T & toUnderType() const { return t; }
 };
 
+
 namespace std
 {
-    template <class T, class Tag>
+    template <typename T, typename Tag>
     struct hash<StrongTypedef<T, Tag>>
     {
         size_t operator()(const StrongTypedef<T, Tag> & x) const

base/common/throwError.h

@@ -1,13 +1,15 @@
 #pragma once
+
 #include <stdexcept>
 
+
 /// Throw DB::Exception-like exception before its definition.
 /// DB::Exception derived from Poco::Exception derived from std::exception.
-/// DB::Exception generally cought as Poco::Exception. std::exception generally has other catch blocks and could lead to other outcomes.
+/// DB::Exception generally caught as Poco::Exception. std::exception generally has other catch blocks and could lead to other outcomes.
 /// DB::Exception is not defined yet. It'd better to throw Poco::Exception but we do not want to include any big header here, even <string>.
 /// So we throw some std::exception instead in the hope its catch block is the same as DB::Exception one.
 template <typename T>
-inline void throwError(const T & err)
+[[noreturn]] inline void throwError(const T & err)
 {
     throw std::runtime_error(err);
 }

base/common/wide_integer.h

@@ -58,9 +58,11 @@ public:
     using signed_base_type = int64_t;
 
     // ctors
-    constexpr integer() noexcept;
+    constexpr integer() noexcept = default;
+
     template <typename T>
     constexpr integer(T rhs) noexcept;
+
     template <typename T>
     constexpr integer(std::initializer_list<T> il) noexcept;
 
@@ -108,9 +110,9 @@ public:
     constexpr explicit operator bool() const noexcept;
 
     template <class T>
-    using __integral_not_wide_integer_class = typename std::enable_if<std::is_arithmetic<T>::value, T>::type;
+    using _integral_not_wide_integer_class = typename std::enable_if<std::is_arithmetic<T>::value, T>::type;
 
-    template <class T, class = __integral_not_wide_integer_class<T>>
+    template <class T, class = _integral_not_wide_integer_class<T>>
     constexpr operator T() const noexcept;
 
     constexpr operator long double() const noexcept;
@@ -119,25 +121,27 @@ public:
 
     struct _impl;
 
+    base_type items[_impl::item_count];
+
 private:
     template <size_t Bits2, typename Signed2>
     friend class integer;
 
     friend class std::numeric_limits<integer<Bits, signed>>;
     friend class std::numeric_limits<integer<Bits, unsigned>>;
-
-    base_type items[_impl::item_count];
 };
 
 template <typename T>
 static constexpr bool ArithmeticConcept() noexcept;
+
 template <class T1, class T2>
-using __only_arithmetic = typename std::enable_if<ArithmeticConcept<T1>() && ArithmeticConcept<T2>()>::type;
+using _only_arithmetic = typename std::enable_if<ArithmeticConcept<T1>() && ArithmeticConcept<T2>()>::type;
 
 template <typename T>
 static constexpr bool IntegralConcept() noexcept;
+
 template <class T, class T2>
-using __only_integer = typename std::enable_if<IntegralConcept<T>() && IntegralConcept<T2>()>::type;
+using _only_integer = typename std::enable_if<IntegralConcept<T>() && IntegralConcept<T2>()>::type;
 
 // Unary operators
 template <size_t Bits, typename Signed>
@@ -153,54 +157,55 @@ constexpr integer<Bits, Signed> operator+(const integer<Bits, Signed> & lhs) noe
 template <size_t Bits, typename Signed, size_t Bits2, typename Signed2>
 std::common_type_t<integer<Bits, Signed>, integer<Bits2, Signed2>> constexpr
 operator*(const integer<Bits, Signed> & lhs, const integer<Bits2, Signed2> & rhs);
-template <typename Arithmetic, typename Arithmetic2, class = __only_arithmetic<Arithmetic, Arithmetic2>>
+template <typename Arithmetic, typename Arithmetic2, class = _only_arithmetic<Arithmetic, Arithmetic2>>
 std::common_type_t<Arithmetic, Arithmetic2> constexpr operator*(const Arithmetic & rhs, const Arithmetic2 & lhs);
 
 template <size_t Bits, typename Signed, size_t Bits2, typename Signed2>
 std::common_type_t<integer<Bits, Signed>, integer<Bits2, Signed2>> constexpr
 operator/(const integer<Bits, Signed> & lhs, const integer<Bits2, Signed2> & rhs);
-template <typename Arithmetic, typename Arithmetic2, class = __only_arithmetic<Arithmetic, Arithmetic2>>
+template <typename Arithmetic, typename Arithmetic2, class = _only_arithmetic<Arithmetic, Arithmetic2>>
 std::common_type_t<Arithmetic, Arithmetic2> constexpr operator/(const Arithmetic & rhs, const Arithmetic2 & lhs);
 
 template <size_t Bits, typename Signed, size_t Bits2, typename Signed2>
 std::common_type_t<integer<Bits, Signed>, integer<Bits2, Signed2>> constexpr
 operator+(const integer<Bits, Signed> & lhs, const integer<Bits2, Signed2> & rhs);
-template <typename Arithmetic, typename Arithmetic2, class = __only_arithmetic<Arithmetic, Arithmetic2>>
+template <typename Arithmetic, typename Arithmetic2, class = _only_arithmetic<Arithmetic, Arithmetic2>>
 std::common_type_t<Arithmetic, Arithmetic2> constexpr operator+(const Arithmetic & rhs, const Arithmetic2 & lhs);
 
 template <size_t Bits, typename Signed, size_t Bits2, typename Signed2>
 std::common_type_t<integer<Bits, Signed>, integer<Bits2, Signed2>> constexpr
 operator-(const integer<Bits, Signed> & lhs, const integer<Bits2, Signed2> & rhs);
-template <typename Arithmetic, typename Arithmetic2, class = __only_arithmetic<Arithmetic, Arithmetic2>>
+template <typename Arithmetic, typename Arithmetic2, class = _only_arithmetic<Arithmetic, Arithmetic2>>
 std::common_type_t<Arithmetic, Arithmetic2> constexpr operator-(const Arithmetic & rhs, const Arithmetic2 & lhs);
 
 template <size_t Bits, typename Signed, size_t Bits2, typename Signed2>
 std::common_type_t<integer<Bits, Signed>, integer<Bits2, Signed2>> constexpr
 operator%(const integer<Bits, Signed> & lhs, const integer<Bits2, Signed2> & rhs);
-template <typename Integral, typename Integral2, class = __only_integer<Integral, Integral2>>
+template <typename Integral, typename Integral2, class = _only_integer<Integral, Integral2>>
 std::common_type_t<Integral, Integral2> constexpr operator%(const Integral & rhs, const Integral2 & lhs);
 
 template <size_t Bits, typename Signed, size_t Bits2, typename Signed2>
 std::common_type_t<integer<Bits, Signed>, integer<Bits2, Signed2>> constexpr
 operator&(const integer<Bits, Signed> & lhs, const integer<Bits2, Signed2> & rhs);
-template <typename Integral, typename Integral2, class = __only_integer<Integral, Integral2>>
+template <typename Integral, typename Integral2, class = _only_integer<Integral, Integral2>>
 std::common_type_t<Integral, Integral2> constexpr operator&(const Integral & rhs, const Integral2 & lhs);
 
 template <size_t Bits, typename Signed, size_t Bits2, typename Signed2>
 std::common_type_t<integer<Bits, Signed>, integer<Bits2, Signed2>> constexpr
 operator|(const integer<Bits, Signed> & lhs, const integer<Bits2, Signed2> & rhs);
-template <typename Integral, typename Integral2, class = __only_integer<Integral, Integral2>>
+template <typename Integral, typename Integral2, class = _only_integer<Integral, Integral2>>
 std::common_type_t<Integral, Integral2> constexpr operator|(const Integral & rhs, const Integral2 & lhs);
 
 template <size_t Bits, typename Signed, size_t Bits2, typename Signed2>
 std::common_type_t<integer<Bits, Signed>, integer<Bits2, Signed2>> constexpr
 operator^(const integer<Bits, Signed> & lhs, const integer<Bits2, Signed2> & rhs);
-template <typename Integral, typename Integral2, class = __only_integer<Integral, Integral2>>
+template <typename Integral, typename Integral2, class = _only_integer<Integral, Integral2>>
 std::common_type_t<Integral, Integral2> constexpr operator^(const Integral & rhs, const Integral2 & lhs);
 
 // TODO: Integral
 template <size_t Bits, typename Signed>
 constexpr integer<Bits, Signed> operator<<(const integer<Bits, Signed> & lhs, int n) noexcept;
+
 template <size_t Bits, typename Signed>
 constexpr integer<Bits, Signed> operator>>(const integer<Bits, Signed> & lhs, int n) noexcept;
 
@@ -217,32 +222,32 @@ constexpr integer<Bits, Signed> operator>>(const integer<Bits, Signed> & lhs, In
 
 template <size_t Bits, typename Signed, size_t Bits2, typename Signed2>
 constexpr bool operator<(const integer<Bits, Signed> & lhs, const integer<Bits2, Signed2> & rhs);
-template <typename Arithmetic, typename Arithmetic2, class = __only_arithmetic<Arithmetic, Arithmetic2>>
+template <typename Arithmetic, typename Arithmetic2, class = _only_arithmetic<Arithmetic, Arithmetic2>>
 constexpr bool operator<(const Arithmetic & rhs, const Arithmetic2 & lhs);
 
 template <size_t Bits, typename Signed, size_t Bits2, typename Signed2>
 constexpr bool operator>(const integer<Bits, Signed> & lhs, const integer<Bits2, Signed2> & rhs);
-template <typename Arithmetic, typename Arithmetic2, class = __only_arithmetic<Arithmetic, Arithmetic2>>
+template <typename Arithmetic, typename Arithmetic2, class = _only_arithmetic<Arithmetic, Arithmetic2>>
 constexpr bool operator>(const Arithmetic & rhs, const Arithmetic2 & lhs);
 
 template <size_t Bits, typename Signed, size_t Bits2, typename Signed2>
 constexpr bool operator<=(const integer<Bits, Signed> & lhs, const integer<Bits2, Signed2> & rhs);
-template <typename Arithmetic, typename Arithmetic2, class = __only_arithmetic<Arithmetic, Arithmetic2>>
+template <typename Arithmetic, typename Arithmetic2, class = _only_arithmetic<Arithmetic, Arithmetic2>>
 constexpr bool operator<=(const Arithmetic & rhs, const Arithmetic2 & lhs);
 
 template <size_t Bits, typename Signed, size_t Bits2, typename Signed2>
 constexpr bool operator>=(const integer<Bits, Signed> & lhs, const integer<Bits2, Signed2> & rhs);
-template <typename Arithmetic, typename Arithmetic2, class = __only_arithmetic<Arithmetic, Arithmetic2>>
+template <typename Arithmetic, typename Arithmetic2, class = _only_arithmetic<Arithmetic, Arithmetic2>>
 constexpr bool operator>=(const Arithmetic & rhs, const Arithmetic2 & lhs);
 
 template <size_t Bits, typename Signed, size_t Bits2, typename Signed2>
 constexpr bool operator==(const integer<Bits, Signed> & lhs, const integer<Bits2, Signed2> & rhs);
-template <typename Arithmetic, typename Arithmetic2, class = __only_arithmetic<Arithmetic, Arithmetic2>>
+template <typename Arithmetic, typename Arithmetic2, class = _only_arithmetic<Arithmetic, Arithmetic2>>
 constexpr bool operator==(const Arithmetic & rhs, const Arithmetic2 & lhs);
 
 template <size_t Bits, typename Signed, size_t Bits2, typename Signed2>
 constexpr bool operator!=(const integer<Bits, Signed> & lhs, const integer<Bits2, Signed2> & rhs);
-template <typename Arithmetic, typename Arithmetic2, class = __only_arithmetic<Arithmetic, Arithmetic2>>
+template <typename Arithmetic, typename Arithmetic2, class = _only_arithmetic<Arithmetic, Arithmetic2>>
 constexpr bool operator!=(const Arithmetic & rhs, const Arithmetic2 & lhs);
 
 }

base/common/wide_integer_impl.h

@@ -5,6 +5,7 @@
 /// (See at http://www.boost.org/LICENSE_1_0.txt)
 
 #include "throwError.h"
+
 #include <cmath>
 #include <cfloat>
 #include <cassert>
@@ -81,7 +82,7 @@ public:
             res.items[T::_impl::big(0)] = std::numeric_limits<typename wide::integer<Bits, Signed>::signed_base_type>::min();
             return res;
         }
-        return 0;
+        return wide::integer<Bits, Signed>(0);
     }
 
     static constexpr wide::integer<Bits, Signed> max() noexcept
@@ -176,7 +177,7 @@ struct integer<Bits, Signed>::_impl
     constexpr static bool is_negative(const integer<B, T> & n) noexcept
     {
         if constexpr (std::is_same_v<T, signed>)
-            return static_cast<signed_base_type>(n.items[big(0)]) < 0;
+            return static_cast<signed_base_type>(n.items[integer<B, T>::_impl::big(0)]) < 0;
         else
             return false;
     }
@@ -193,40 +194,36 @@ struct integer<Bits, Signed>::_impl
     template <size_t B, class S>
     constexpr static integer<B, S> make_positive(const integer<B, S> & n) noexcept
     {
-        return is_negative(n) ? operator_unary_minus(n) : n;
+        return is_negative(n) ? integer<B, S>(operator_unary_minus(n)) : n;
     }
 
     template <typename T>
     __attribute__((no_sanitize("undefined"))) constexpr static auto to_Integral(T f) noexcept
     {
-        if constexpr (std::is_same_v<T, __int128>)
-            return f;
-        else if constexpr (std::is_signed_v<T>)
+        if constexpr (std::is_signed_v<T>)
             return static_cast<int64_t>(f);
         else
             return static_cast<uint64_t>(f);
     }
 
     template <typename Integral>
-    constexpr static void wide_integer_from_bultin(integer<Bits, Signed> & self, Integral rhs) noexcept
+    constexpr static void wide_integer_from_builtin(integer<Bits, Signed> & self, Integral rhs) noexcept
     {
-        self.items[0] = _impl::to_Integral(rhs);
-        if constexpr (std::is_same_v<Integral, __int128>)
-            self.items[1] = rhs >> base_bits;
+        static_assert(sizeof(Integral) <= sizeof(base_type));
 
-        constexpr const unsigned start = (sizeof(Integral) == 16) ? 2 : 1;
+        self.items[0] = _impl::to_Integral(rhs);
 
         if constexpr (std::is_signed_v<Integral>)
         {
             if (rhs < 0)
             {
-                for (unsigned i = start; i < item_count; ++i)
+                for (size_t i = 1; i < item_count; ++i)
                     self.items[i] = -1;
                 return;
             }
         }
 
-        for (unsigned i = start; i < item_count; ++i)
+        for (size_t i = 1; i < item_count; ++i)
             self.items[i] = 0;
     }
 
@@ -239,7 +236,8 @@ struct integer<Bits, Signed>::_impl
      * a_(n - 1) = a_n * max_int + b2, a_n <= max_int <- base case.
      */
     template <class T>
-    constexpr static void set_multiplier(integer<Bits, Signed> & self, T t) noexcept {
+    constexpr static void set_multiplier(integer<Bits, Signed> & self, T t) noexcept
+    {
         constexpr uint64_t max_int = std::numeric_limits<uint64_t>::max();
 
         /// Implementation specific behaviour on overflow (if we don't check here, stack overflow will triggered in bigint_cast).
@@ -260,7 +258,8 @@ struct integer<Bits, Signed>::_impl
         self += static_cast<uint64_t>(t - alpha * static_cast<T>(max_int)); // += b_i
     }
 
-    constexpr static void wide_integer_from_bultin(integer<Bits, Signed>& self, double rhs) noexcept {
+    constexpr static void wide_integer_from_builtin(integer<Bits, Signed>& self, double rhs) noexcept
+    {
         constexpr int64_t max_int = std::numeric_limits<int64_t>::max();
         constexpr int64_t min_int = std::numeric_limits<int64_t>::min();
 
@@ -383,13 +382,13 @@ struct integer<Bits, Signed>::_impl
             if (bit_shift)
                 lhs.items[big(items_shift)] |= std::numeric_limits<base_type>::max() << (base_bits - bit_shift);
 
-            for (unsigned i = item_count - items_shift; i < items_shift; ++i)
-                lhs.items[little(i)] = std::numeric_limits<base_type>::max();
+            for (unsigned i = 0; i < items_shift; ++i)
+                lhs.items[big(i)] = std::numeric_limits<base_type>::max();
         }
         else
         {
-            for (unsigned i = item_count - items_shift; i < items_shift; ++i)
-                lhs.items[little(i)] = 0;
+            for (unsigned i = 0; i < items_shift; ++i)
+                lhs.items[big(i)] = 0;
         }
 
         return lhs;
@@ -397,23 +396,23 @@ struct integer<Bits, Signed>::_impl
 
 private:
     template <typename T>
-    constexpr static base_type get_item(const T & x, unsigned number)
+    constexpr static base_type get_item(const T & x, unsigned idx)
     {
         if constexpr (IsWideInteger<T>::value)
         {
-            if (number < T::_impl::item_count)
-                return x.items[number];
+            if (idx < T::_impl::item_count)
+                return x.items[idx];
             return 0;
         }
         else
         {
             if constexpr (sizeof(T) <= sizeof(base_type))
             {
-                if (!number)
+                if (0 == idx)
                     return x;
             }
-            else if (number * sizeof(base_type) < sizeof(T))
-                return x >> (number * base_bits); // & std::numeric_limits<base_type>::max()
+            else if (idx * sizeof(base_type) < sizeof(T))
+                return x >> (idx * base_bits); // & std::numeric_limits<base_type>::max()
             return 0;
         }
     }
@@ -439,7 +438,7 @@ private:
 
         for (unsigned i = 1; i < item_count; ++i)
         {
-            if (underflows[i-1])
+            if (underflows[i - 1])
             {
                 base_type & res_item = res.items[little(i)];
                 if (res_item == 0)
@@ -472,7 +471,7 @@ private:
 
         for (unsigned i = 1; i < item_count; ++i)
         {
-            if (overflows[i-1])
+            if (overflows[i - 1])
             {
                 base_type & res_item = res.items[little(i)];
                 ++res_item;
@@ -532,6 +531,17 @@ private:
             res.items[little(2)] = r12 >> 64;
             return res;
         }
+        else if constexpr (Bits == 128 && sizeof(base_type) == 8)
+        {
+            using CompilerUInt128 = unsigned __int128;
+            CompilerUInt128 a = (CompilerUInt128(lhs.items[1]) << 64) + lhs.items[0];
+            CompilerUInt128 b = (CompilerUInt128(rhs.items[1]) << 64) + rhs.items[0];
+            CompilerUInt128 c = a * b;
+            integer<Bits, Signed> res;
+            res.items[0] = c;
+            res.items[1] = c >> 64;
+            return res;
+        }
         else
         {
             integer<Bits, Signed> res{};
@@ -657,7 +667,7 @@ public:
     }
 
     template <typename T>
-    constexpr static bool operator_more(const integer<Bits, Signed> & lhs, const T & rhs) noexcept
+    constexpr static bool operator_greater(const integer<Bits, Signed> & lhs, const T & rhs) noexcept
     {
         if constexpr (should_keep_size<T>())
         {
@@ -677,7 +687,7 @@ public:
         else
         {
             static_assert(IsWideInteger<T>::value);
-            return std::common_type_t<integer<Bits, Signed>, T>::_impl::operator_more(T(lhs), rhs);
+            return std::common_type_t<integer<Bits, Signed>, T>::_impl::operator_greater(T(lhs), rhs);
         }
     }
 
@@ -764,7 +774,6 @@ public:
         }
     }
 
-private:
     template <typename T>
     constexpr static bool is_zero(const T & x)
     {
@@ -781,46 +790,65 @@ private:
     }
 
     /// returns quotient as result and remainder in numerator.
-    template <typename T>
-    constexpr static T divide(T & numerator, T && denominator)
+    template <size_t Bits2>
+    constexpr static integer<Bits2, unsigned> divide(integer<Bits2, unsigned> & numerator, integer<Bits2, unsigned> denominator)
     {
-        if (is_zero(denominator))
-            throwError("divide by zero");
+        static_assert(std::is_unsigned_v<Signed>);
 
-        T & n = numerator;
-        T & d = denominator;
-        T x = 1;
-        T quotient = 0;
-
-        while (!operator_more(d, n) && operator_eq(operator_amp(shift_right(d, base_bits * item_count - 1), 1), 0))
+        if constexpr (Bits == 128 && sizeof(base_type) == 8)
         {
-            x = shift_left(x, 1);
-            d = shift_left(d, 1);
+            using CompilerUInt128 = unsigned __int128;
+
+            CompilerUInt128 a = (CompilerUInt128(numerator.items[1]) << 64) + numerator.items[0];
+            CompilerUInt128 b = (CompilerUInt128(denominator.items[1]) << 64) + denominator.items[0];
+            CompilerUInt128 c = a / b;
+
+            integer<Bits, Signed> res;
+            res.items[0] = c;
+            res.items[1] = c >> 64;
+
+            CompilerUInt128 remainder = a - b * c;
+            numerator.items[0] = remainder;
+            numerator.items[1] = remainder >> 64;
+
+            return res;
         }
 
-        while (!operator_eq(x, 0))
+        if (is_zero(denominator))
+            throwError("Division by zero");
+
+        integer<Bits2, unsigned> x = 1;
+        integer<Bits2, unsigned> quotient = 0;
+
+        while (!operator_greater(denominator, numerator) && is_zero(operator_amp(shift_right(denominator, Bits2 - 1), 1)))
         {
-            if (!operator_more(d, n))
+            x = shift_left(x, 1);
+            denominator = shift_left(denominator, 1);
+        }
+
+        while (!is_zero(x))
+        {
+            if (!operator_greater(denominator, numerator))
             {
-                n = operator_minus(n, d);
+                numerator = operator_minus(numerator, denominator);
                 quotient = operator_pipe(quotient, x);
             }
 
             x = shift_right(x, 1);
-            d = shift_right(d, 1);
+            denominator = shift_right(denominator, 1);
         }
 
         return quotient;
     }
 
-public:
     template <typename T>
     constexpr static auto operator_slash(const integer<Bits, Signed> & lhs, const T & rhs)
     {
         if constexpr (should_keep_size<T>())
         {
-            integer<Bits, Signed> numerator = make_positive(lhs);
-            integer<Bits, Signed> quotient = divide(numerator, make_positive(integer<Bits, Signed>(rhs)));
+            integer<Bits, unsigned> numerator = make_positive(lhs);
+            integer<Bits, unsigned> denominator = make_positive(integer<Bits, Signed>(rhs));
+            integer<Bits, unsigned> quotient = integer<Bits, unsigned>::_impl::divide(numerator, std::move(denominator));
 
             if (std::is_same_v<Signed, signed> && is_negative(rhs) != is_negative(lhs))
                 quotient = operator_unary_minus(quotient);
@@ -838,8 +866,9 @@ public:
     {
         if constexpr (should_keep_size<T>())
         {
-            integer<Bits, Signed> remainder = make_positive(lhs);
-            divide(remainder, make_positive(integer<Bits, Signed>(rhs)));
+            integer<Bits, unsigned> remainder = make_positive(lhs);
+            integer<Bits, unsigned> denominator = make_positive(integer<Bits, Signed>(rhs));
+            integer<Bits, unsigned>::_impl::divide(remainder, std::move(denominator));
 
             if (std::is_same_v<Signed, signed> && is_negative(lhs))
                 remainder = operator_unary_minus(remainder);
@@ -905,7 +934,7 @@ public:
                     ++c;
                 }
                 else
-                    throwError("invalid char from");
+                    throwError("Invalid char from");
             }
         }
         else
@@ -913,7 +942,7 @@ public:
             while (*c)
             {
                 if (*c < '0' || *c > '9')
-                    throwError("invalid char from");
+                    throwError("Invalid char from");
 
                 res = multiply(res, 10U);
                 res = plus(res, *c - '0');
@@ -930,11 +959,6 @@ public:
 
 // Members
 
-template <size_t Bits, typename Signed>
-constexpr integer<Bits, Signed>::integer() noexcept
-    : items{}
-{}
-
 template <size_t Bits, typename Signed>
 template <typename T>
 constexpr integer<Bits, Signed>::integer(T rhs) noexcept
@@ -943,7 +967,7 @@ constexpr integer<Bits, Signed>::integer(T rhs) noexcept
     if constexpr (IsWideInteger<T>::value)
         _impl::wide_integer_from_wide_integer(*this, rhs);
     else
-        _impl::wide_integer_from_bultin(*this, rhs);
+        _impl::wide_integer_from_builtin(*this, rhs);
 }
 
 template <size_t Bits, typename Signed>
@@ -956,10 +980,19 @@ constexpr integer<Bits, Signed>::integer(std::initializer_list<T> il) noexcept
         if constexpr (IsWideInteger<T>::value)
             _impl::wide_integer_from_wide_integer(*this, *il.begin());
         else
-            _impl::wide_integer_from_bultin(*this, *il.begin());
+            _impl::wide_integer_from_builtin(*this, *il.begin());
+    }
+    else if (il.size() == 0)
+    {
+        _impl::wide_integer_from_builtin(*this, 0);
     }
     else
-        _impl::wide_integer_from_bultin(*this, 0);
+    {
+        auto it = il.begin();
+        for (size_t i = 0; i < _impl::item_count; ++i)
+            if (it < il.end())
+                items[i] = *it;
+    }
 }
 
 template <size_t Bits, typename Signed>
@@ -974,7 +1007,7 @@ template <size_t Bits, typename Signed>
 template <typename T>
 constexpr integer<Bits, Signed> & integer<Bits, Signed>::operator=(T rhs) noexcept
 {
-    _impl::wide_integer_from_bultin(*this, rhs);
+    _impl::wide_integer_from_builtin(*this, rhs);
     return *this;
 }
 
@@ -1057,7 +1090,7 @@ constexpr integer<Bits, Signed> & integer<Bits, Signed>::operator>>=(int n) noex
 {
     if (static_cast<size_t>(n) >= Bits)
     {
-        if (is_negative(*this))
+        if (_impl::is_negative(*this))
             *this = -1;
         else
             *this = 0;
@@ -1107,16 +1140,17 @@ template <size_t Bits, typename Signed>
 template <class T, class>
 constexpr integer<Bits, Signed>::operator T() const noexcept
 {
-    if constexpr (std::is_same_v<T, __int128>)
-    {
-        static_assert(Bits >= 128);
-        return (__int128(items[1]) << 64) | items[0];
-    }
-    else
-    {
-        static_assert(std::numeric_limits<T>::is_integer);
-        return items[0];
-    }
+    static_assert(std::numeric_limits<T>::is_integer);
+
+    /// NOTE: memcpy will suffice, but unfortunately, this function is constexpr.
+
+    using UnsignedT = std::make_unsigned_t<T>;
+
+    UnsignedT res{};
+    for (unsigned i = 0; i < _impl::item_count && i < (sizeof(T) + sizeof(base_type) - 1) / sizeof(base_type); ++i)
+        res += UnsignedT(items[i]) << (sizeof(base_type) * 8 * i);
+
+    return res;
 }
 
 template <size_t Bits, typename Signed>
@@ -1280,7 +1314,7 @@ template <size_t Bits, typename Signed>
 constexpr integer<Bits, Signed> operator<<(const integer<Bits, Signed> & lhs, int n) noexcept
 {
     if (static_cast<size_t>(n) >= Bits)
-        return 0;
+        return integer<Bits, Signed>(0);
     if (n <= 0)
         return lhs;
     return integer<Bits, Signed>::_impl::shift_left(lhs, n);
@@ -1289,7 +1323,7 @@ template <size_t Bits, typename Signed>
 constexpr integer<Bits, Signed> operator>>(const integer<Bits, Signed> & lhs, int n) noexcept
 {
     if (static_cast<size_t>(n) >= Bits)
-        return 0;
+        return integer<Bits, Signed>(0);
     if (n <= 0)
         return lhs;
     return integer<Bits, Signed>::_impl::shift_right(lhs, n);
@@ -1309,7 +1343,7 @@ constexpr bool operator<(const Arithmetic & lhs, const Arithmetic2 & rhs)
 template <size_t Bits, typename Signed, size_t Bits2, typename Signed2>
 constexpr bool operator>(const integer<Bits, Signed> & lhs, const integer<Bits2, Signed2> & rhs)
 {
-    return std::common_type_t<integer<Bits, Signed>, integer<Bits2, Signed2>>::_impl::operator_more(lhs, rhs);
+    return std::common_type_t<integer<Bits, Signed>, integer<Bits2, Signed2>>::_impl::operator_greater(lhs, rhs);
 }
 template <typename Arithmetic, typename Arithmetic2, class>
 constexpr bool operator>(const Arithmetic & lhs, const Arithmetic2 & rhs)
@@ -1332,7 +1366,7 @@ constexpr bool operator<=(const Arithmetic & lhs, const Arithmetic2 & rhs)
 template <size_t Bits, typename Signed, size_t Bits2, typename Signed2>
 constexpr bool operator>=(const integer<Bits, Signed> & lhs, const integer<Bits2, Signed2> & rhs)
 {
-    return std::common_type_t<integer<Bits, Signed>, integer<Bits2, Signed2>>::_impl::operator_more(lhs, rhs)
+    return std::common_type_t<integer<Bits, Signed>, integer<Bits2, Signed2>>::_impl::operator_greater(lhs, rhs)
         || std::common_type_t<integer<Bits, Signed>, integer<Bits2, Signed2>>::_impl::operator_eq(lhs, rhs);
 }
 template <typename Arithmetic, typename Arithmetic2, class>

base/common/wide_integer_to_string.h

@@ -1,9 +1,12 @@
 #pragma once
 
 #include <string>
+#include <ostream>
+#include <fmt/format.h>
 
 #include "wide_integer.h"
 
+
 namespace wide
 {
 
@@ -33,3 +36,34 @@ inline std::string to_string(const integer<Bits, Signed> & n)
 }
 
 }
+
+
+template <size_t Bits, typename Signed>
+std::ostream & operator<<(std::ostream & out, const wide::integer<Bits, Signed> & value)
+{
+    return out << to_string(value);
+}
+
+
+/// See https://fmt.dev/latest/api.html#formatting-user-defined-types
+template <size_t Bits, typename Signed>
+struct fmt::formatter<wide::integer<Bits, Signed>>
+{
+    constexpr auto parse(format_parse_context & ctx)
+    {
+        auto it = ctx.begin();
+        auto end = ctx.end();
+
+        /// Only support {}.
+        if (it != end && *it != '}')
+            throw format_error("invalid format");
+
+        return it;
+    }
+
+    template <typename FormatContext>
+    auto format(const wide::integer<Bits, Signed> & value, FormatContext & ctx)
+    {
+        return format_to(ctx.out(), "{}", to_string(value));
+    }
+};

contrib/datasketches-cpp

@@ -1 +1 @@
-Subproject commit f915d35b2de676683493c86c585141a1e1c83334
+Subproject commit 7d73d7610db31d4e1ecde0fb3a7ee90ef371207f

docker/test/fasttest/run.sh

@@ -308,6 +308,8 @@ function run_tests
         01354_order_by_tuple_collate_const
         01355_ilike
         01411_bayesian_ab_testing
+        01798_uniq_theta_sketch
+        01799_long_uniq_theta_sketch
         collate
         collation
         _orc_

docs/en/sql-reference/aggregate-functions/reference/uniq.md

@@ -38,3 +38,4 @@ We recommend using this function in almost all scenarios.
 -   [uniqCombined64](../../../sql-reference/aggregate-functions/reference/uniqcombined64.md#agg_function-uniqcombined64)
 -   [uniqHLL12](../../../sql-reference/aggregate-functions/reference/uniqhll12.md#agg_function-uniqhll12)
 -   [uniqExact](../../../sql-reference/aggregate-functions/reference/uniqexact.md#agg_function-uniqexact)
+-   [uniqThetaSketch](../../../sql-reference/aggregate-functions/reference/uniqthetasketch.md#agg_function-uniqthetasketch)

docs/en/sql-reference/aggregate-functions/reference/uniqcombined.md

@@ -49,3 +49,4 @@ Compared to the [uniq](../../../sql-reference/aggregate-functions/reference/uniq
 -   [uniqCombined64](../../../sql-reference/aggregate-functions/reference/uniqcombined64.md#agg_function-uniqcombined64)
 -   [uniqHLL12](../../../sql-reference/aggregate-functions/reference/uniqhll12.md#agg_function-uniqhll12)
 -   [uniqExact](../../../sql-reference/aggregate-functions/reference/uniqexact.md#agg_function-uniqexact)
+-   [uniqThetaSketch](../../../sql-reference/aggregate-functions/reference/uniqthetasketch.md#agg_function-uniqthetasketch)

docs/en/sql-reference/aggregate-functions/reference/uniqexact.md

@@ -23,3 +23,4 @@ The function takes a variable number of parameters. Parameters can be `Tuple`, `
 -   [uniq](../../../sql-reference/aggregate-functions/reference/uniq.md#agg_function-uniq)
 -   [uniqCombined](../../../sql-reference/aggregate-functions/reference/uniq.md#agg_function-uniqcombined)
 -   [uniqHLL12](../../../sql-reference/aggregate-functions/reference/uniq.md#agg_function-uniqhll12)
+-   [uniqThetaSketch](../../../sql-reference/aggregate-functions/reference/uniqthetasketch.md#agg_function-uniqthetasketch)

docs/en/sql-reference/aggregate-functions/reference/uniqhll12.md

@@ -37,3 +37,4 @@ We don’t recommend using this function. In most cases, use the [uniq](../../..
 -   [uniq](../../../sql-reference/aggregate-functions/reference/uniq.md#agg_function-uniq)
 -   [uniqCombined](../../../sql-reference/aggregate-functions/reference/uniqcombined.md#agg_function-uniqcombined)
 -   [uniqExact](../../../sql-reference/aggregate-functions/reference/uniqexact.md#agg_function-uniqexact)
+-   [uniqThetaSketch](../../../sql-reference/aggregate-functions/reference/uniqthetasketch.md#agg_function-uniqthetasketch)

docs/en/sql-reference/aggregate-functions/reference/uniqthetasketch.md

@@ -0,0 +1,39 @@
+---
+toc_priority: 195
+---
+
+# uniqThetaSketch {#agg_function-uniqthetasketch}
+
+Calculates the approximate number of different argument values, using the [Theta Sketch Framework](https://datasketches.apache.org/docs/Theta/ThetaSketchFramework.html).
+
+``` sql
+uniqThetaSketch(x[, ...])
+```
+
+**Arguments**
+
+The function takes a variable number of parameters. Parameters can be `Tuple`, `Array`, `Date`, `DateTime`, `String`, or numeric types.
+
+**Returned value**
+
+-   A [UInt64](../../../sql-reference/data-types/int-uint.md)-type number.
+
+**Implementation details**
+
+Function:
+
+-   Calculates a hash for all parameters in the aggregate, then uses it in calculations.
+
+-   Uses the [KMV](https://datasketches.apache.org/docs/Theta/InverseEstimate.html) algorithm to approximate the number of different argument values.
+
+        4096(2^12) 64-bit sketch are used. The size of the state is about 41 KB.
+
+-   The relative error is 3.125% (95% confidence), see the [relative error table](https://datasketches.apache.org/docs/Theta/ThetaErrorTable.html) for detail.
+
+**See Also**
+
+-   [uniq](../../../sql-reference/aggregate-functions/reference/uniq.md#agg_function-uniq)
+-   [uniqCombined](../../../sql-reference/aggregate-functions/reference/uniqcombined.md#agg_function-uniqcombined)
+-   [uniqCombined64](../../../sql-reference/aggregate-functions/reference/uniqcombined64.md#agg_function-uniqcombined64)
+-   [uniqHLL12](../../../sql-reference/aggregate-functions/reference/uniqhll12.md#agg_function-uniqhll12)
+-   [uniqExact](../../../sql-reference/aggregate-functions/reference/uniqexact.md#agg_function-uniqexact)

docs/en/sql-reference/dictionaries/external-dictionaries/external-dicts-dict-layout.md

@@ -95,7 +95,9 @@ LAYOUT(FLAT(INITIAL_ARRAY_SIZE 50000 MAX_ARRAY_SIZE 5000000))
 
 The dictionary is completely stored in memory in the form of a hash table. The dictionary can contain any number of elements with any identifiers In practice, the number of keys can reach tens of millions of items.
 
-The hash table will be preallocated (this will make dictionary load faster), if the is approx number of total rows is known, this is supported only if the source is `clickhouse` without any `<where>` (since in case of `<where>` you can filter out too much rows and the dictionary will allocate too much memory, that will not be used eventually).
+If `preallocate` is `true` (default is `false`) the hash table will be preallocated (this will make dictionary load faster). But note that you should use it only if:
+- the source support approximate number of elements (for now it is supported only by the `ClickHouse` source)
+- there is no duplicates in the data (otherwise it may increase memory usage for the hashtable)
 
 All types of sources are supported. When updating, data (from a file or from a table) is read in its entirety.
 
@@ -103,21 +105,23 @@ Configuration example:
 
 ``` xml
 <layout>
-  <hashed />
+  <hashed>
+    <preallocate>0</preallocate>
+  </hashed>
 </layout>
 ```
 
 or
 
 ``` sql
-LAYOUT(HASHED())
+LAYOUT(HASHED(PREALLOCATE 0))
 ```
 
 ### sparse_hashed {#dicts-external_dicts_dict_layout-sparse_hashed}
 
 Similar to `hashed`, but uses less memory in favor more CPU usage.
 
-It will be also preallocated so as `hashed`, note that it is even more significant for `sparse_hashed`.
+It will be also preallocated so as `hashed` (with `preallocate` set to `true`), and note that it is even more significant for `sparse_hashed`.
 
 Configuration example:
 
@@ -127,8 +131,10 @@ Configuration example:
 </layout>
 ```
 
+or
+
 ``` sql
-LAYOUT(SPARSE_HASHED())
+LAYOUT(SPARSE_HASHED([PREALLOCATE 0]))
 ```
 
 ### complex_key_hashed {#complex-key-hashed}

docs/en/sql-reference/functions/array-functions.md

@@ -1544,3 +1544,52 @@ SELECT arrayCumSumNonNegative([1, 1, -4, 1]) AS res
 ```
 Note that the `arraySumNonNegative` is a [higher-order function](../../sql-reference/functions/index.md#higher-order-functions). You can pass a lambda function to it as the first argument.
 
+## arrayProduct {#arrayproduct}
+
+Multiplies elements of an [array](../../sql-reference/data-types/array.md).
+
+**Syntax**
+
+``` sql
+arrayProduct(arr)
+```
+
+**Arguments**
+
+-   `arr` — [Array](../../sql-reference/data-types/array.md) of numeric values.
+
+**Returned value**
+
+-   A product of array's elements.
+
+Type: [Float64](../../sql-reference/data-types/float.md).
+
+**Examples**
+
+Query:
+
+``` sql
+SELECT arrayProduct([1,2,3,4,5,6]) as res;
+```
+
+Result:
+
+``` text
+┌─res───┐
+│ 720   │
+└───────┘
+```
+
+Query:
+
+``` sql
+SELECT arrayProduct([toDecimal64(1,8), toDecimal64(2,8), toDecimal64(3,8)]) as res, toTypeName(res);
+```
+
+Return value type is always [Float64](../../sql-reference/data-types/float.md). Result:
+
+``` text
+┌─res─┬─toTypeName(arrayProduct(array(toDecimal64(1, 8), toDecimal64(2, 8), toDecimal64(3, 8))))─┐
+│ 6   │ Float64                                                                                  │
+└─────┴──────────────────────────────────────────────────────────────────────────────────────────┘
+```

docs/en/sql-reference/statements/explain.md

@@ -47,9 +47,9 @@ Union
 
 ### EXPLAIN AST {#explain-ast}
 
-Dump query AST.
+Dump query AST. Supports all types of queries, not only `SELECT`.
 
-Example:
+Examples:
 
 ```sql
 EXPLAIN AST SELECT 1;
@@ -63,6 +63,22 @@ SelectWithUnionQuery (children 1)
     Literal UInt64_1
 ```
 
+```sql
+EXPLAIN AST ALTER TABLE t1 DELETE WHERE date = today();
+```
+
+```sql
+  explain
+  AlterQuery  t1 (children 1)
+   ExpressionList (children 1)
+    AlterCommand 27 (children 1)
+     Function equals (children 1)
+      ExpressionList (children 2)
+       Identifier date
+       Function today (children 1)
+        ExpressionList
+```
+
 ### EXPLAIN SYNTAX {#explain-syntax}
 
 Returns query after syntax optimizations.

docs/ru/engines/table-engines/integrations/s3.md

@@ -105,7 +105,7 @@ SELECT * FROM s3_engine_table LIMIT 2;
 
 ## Примеры использования {#usage-examples}
 
-Предположим, у нас есть несколько файлов в формате TSV со следующими URL-адресами в HDFS:
+Предположим, у нас есть несколько файлов в формате TSV со следующими URL-адресами в S3:
 
 -   'https://storage.yandexcloud.net/my-test-bucket-768/some_prefix/some_file_1.csv'
 -   'https://storage.yandexcloud.net/my-test-bucket-768/some_prefix/some_file_2.csv'

docs/ru/operations/tips.md

@@ -1,248 +0,0 @@
----
-toc_priority: 58
-toc_title: "Советы по эксплуатации"
----
-
-# Советы по эксплуатации {#sovety-po-ekspluatatsii}
-
-## CPU Scaling Governor {#cpu-scaling-governor}
-
-Всегда используйте `performance` scaling governor. `ondemand` scaling governor работает намного хуже при постоянно высоком спросе.
-
-``` bash
-$ echo 'performance' | sudo tee /sys/devices/system/cpu/cpu*/cpufreq/scaling_governor
-```
-
-## Ограничение CPU {#ogranichenie-cpu}
-
-Процессоры могут перегреваться. С помощью `dmesg` можно увидеть, если тактовая частота процессора была ограничена из-за перегрева.
-Также ограничение может устанавливаться снаружи на уровне дата-центра. С помощью `turbostat` можно за этим наблюдать под нагрузкой.
-
-## Оперативная память {#operativnaia-pamiat}
-
-Для небольших объёмов данных (до ~200 Гб в сжатом виде) лучше всего использовать столько памяти не меньше, чем объём данных.
-Для больших объёмов данных, при выполнении интерактивных (онлайн) запросов, стоит использовать разумный объём оперативной памяти (128 Гб или более) для того, чтобы горячее подмножество данных поместилось в кеше страниц.
-Даже для объёмов данных в ~50 Тб на сервер, использование 128 Гб оперативной памяти намного лучше для производительности выполнения запросов, чем 64 Гб.
-
-Не выключайте overcommit. Значение `cat /proc/sys/vm/overcommit_memory` должно быть 0 or 1. Выполните:
-
-``` bash
-$ echo 0 | sudo tee /proc/sys/vm/overcommit_memory
-```
-
-## Huge Pages {#huge-pages}
-
-Механизм прозрачных huge pages нужно отключить. Он мешает работе аллокаторов памяти, что приводит к значительной деградации производительности.
-
-``` bash
-$ echo 'madvise' | sudo tee /sys/kernel/mm/transparent_hugepage/enabled
-```
-
-С помощью `perf top` можно наблюдать за временем, проведенном в ядре операционной системы для управления памятью.
-Постоянные huge pages так же не нужно аллоцировать.
-
-## Подсистема хранения {#podsistema-khraneniia}
-
-Если ваш бюджет позволяет использовать SSD, используйте SSD.
-В противном случае используйте HDD. SATA HDDs 7200 RPM подойдут.
-
-Предпочитайте много серверов с локальными жесткими дисками вместо меньшего числа серверов с подключенными дисковыми полками.
-Но для хранения архивов с редкими запросами полки всё же подходят.
-
-## RAID {#raid}
-
-При использовании HDD можно объединить их RAID-10, RAID-5, RAID-6 или RAID-50.
-Лучше использовать программный RAID в Linux (`mdadm`). Лучше не использовать LVM.
-При создании RAID-10, нужно выбрать `far` расположение.
-Если бюджет позволяет, лучше выбрать RAID-10.
-
-На более чем 4 дисках вместо RAID-5 нужно использовать RAID-6 (предпочтительнее) или RAID-50.
-При использовании RAID-5, RAID-6 или RAID-50, нужно всегда увеличивать stripe_cache_size, так как значение по умолчанию выбрано не самым удачным образом.
-
-``` bash
-$ echo 4096 | sudo tee /sys/block/md2/md/stripe_cache_size
-```
-
-Точное число стоит вычислять из числа устройств и размер блока по формуле: `2 * num_devices * chunk_size_in_bytes / 4096`.
-
-Размер блока в 1024 Кб подходит для всех конфигураций RAID.
-Никогда не указывайте слишком маленький или слишком большой размер блока.
-
-На SSD можно использовать RAID-0.
-Вне зависимости от использования RAID, всегда используйте репликацию для безопасности данных.
-
-Включите NCQ с длинной очередью. Для HDD стоит выбрать планировщик CFQ, а для SSD — noop. Не стоит уменьшать настройку readahead.
-На HDD стоит включать кеш записи.
-
-## Файловая система {#failovaia-sistema}
-
-Ext4 самый проверенный вариант. Укажите опции монтирования `noatime,nobarrier`.
-XFS также подходит, но не так тщательно протестирована в сочетании с ClickHouse.
-Большинство других файловых систем также должны нормально работать. Файловые системы с отложенной аллокацией работают лучше.
-
-## Ядро Linux {#iadro-linux}
-
-Не используйте слишком старое ядро Linux.
-
-## Сеть {#set}
-
-При использовании IPv6, стоит увеличить размер кеша маршрутов.
-Ядра Linux до 3.2 имели массу проблем в реализации IPv6.
-
-Предпочитайте как минимум 10 Гбит сеть. 1 Гбит также будет работать, но намного хуже для починки реплик с десятками терабайт данных или для обработки распределенных запросов с большим объёмом промежуточных данных.
-
-## ZooKeeper {#zookeeper}
-
-Вероятно вы уже используете ZooKeeper для других целей. Можно использовать ту же инсталляцию ZooKeeper, если она не сильно перегружена.
-
-Лучше использовать свежую версию ZooKeeper, как минимум 3.4.9. Версия в стабильных дистрибутивах Linux может быть устаревшей.
-
-Никогда не используете написанные вручную скрипты для переноса данных между разными ZooKeeper кластерами, потому что результат будет некорректный для sequential нод. Никогда не используйте утилиту «zkcopy», по той же причине: https://github.com/ksprojects/zkcopy/issues/15
-
-Если вы хотите разделить существующий ZooKeeper кластер на два, правильный способ - увеличить количество его реплик, а затем переконфигурировать его как два независимых кластера.
-
-Не запускайте ZooKeeper на тех же серверах, что и ClickHouse. Потому что ZooKeeper очень чувствителен к задержкам, а ClickHouse может использовать все доступные системные ресурсы.
-
-С настройками по умолчанию, ZooKeeper является бомбой замедленного действия:
-
-> Сервер ZooKeeper не будет удалять файлы со старыми снепшоты и логами при использовании конфигурации по умолчанию (см. autopurge), это является ответственностью оператора.
-
-Эту бомбу нужно обезвредить.
-
-Далее описана конфигурация ZooKeeper (3.5.1), используемая в боевом окружении Яндекс.Метрики на момент 20 мая 2017 года:
-
-zoo.cfg:
-
-``` bash
-# http://hadoop.apache.org/zookeeper/docs/current/zookeeperAdmin.html
-
-# The number of milliseconds of each tick
-tickTime=2000
-# The number of ticks that the initial
-# synchronization phase can take
-initLimit=30000
-# The number of ticks that can pass between
-# sending a request and getting an acknowledgement
-syncLimit=10
-
-maxClientCnxns=2000
-
-maxSessionTimeout=60000000
-# the directory where the snapshot is stored.
-dataDir=/opt/zookeeper/{{ '{{' }} cluster['name'] {{ '{{' }} '}}' }}/data
-# Place the dataLogDir to a separate physical disc for better performance
-dataLogDir=/opt/zookeeper/{{ '{{' }} cluster['name'] {{ '{{' }} '}}' }}/logs
-
-autopurge.snapRetainCount=10
-autopurge.purgeInterval=1
-
-
-# To avoid seeks ZooKeeper allocates space in the transaction log file in
-# blocks of preAllocSize kilobytes. The default block size is 64M. One reason
-# for changing the size of the blocks is to reduce the block size if snapshots
-# are taken more often. (Also, see snapCount).
-preAllocSize=131072
-
-# Clients can submit requests faster than ZooKeeper can process them,
-# especially if there are a lot of clients. To prevent ZooKeeper from running
-# out of memory due to queued requests, ZooKeeper will throttle clients so that
-# there is no more than globalOutstandingLimit outstanding requests in the
-# system. The default limit is 1,000.ZooKeeper logs transactions to a
-# transaction log. After snapCount transactions are written to a log file a
-# snapshot is started and a new transaction log file is started. The default
-# snapCount is 10,000.
-snapCount=3000000
-
-# If this option is defined, requests will be will logged to a trace file named
-# traceFile.year.month.day.
-#traceFile=
-
-# Leader accepts client connections. Default value is "yes". The leader machine
-# coordinates updates. For higher update throughput at thes slight expense of
-# read throughput the leader can be configured to not accept clients and focus
-# on coordination.
-leaderServes=yes
-
-standaloneEnabled=false
-dynamicConfigFile=/etc/zookeeper-{{ '{{' }} cluster['name'] {{ '{{' }} '}}' }}/conf/zoo.cfg.dynamic
-```
-
-Версия Java:
-
-``` text
-Java(TM) SE Runtime Environment (build 1.8.0_25-b17)
-Java HotSpot(TM) 64-Bit Server VM (build 25.25-b02, mixed mode)
-```
-
-Параметры JVM:
-
-``` bash
-NAME=zookeeper-{{ '{{' }} cluster['name'] {{ '{{' }} '}}' }}
-ZOOCFGDIR=/etc/$NAME/conf
-
-# TODO this is really ugly
-# How to find out, which jars are needed?
-# seems, that log4j requires the log4j.properties file to be in the classpath
-CLASSPATH="$ZOOCFGDIR:/usr/build/classes:/usr/build/lib/*.jar:/usr/share/zookeeper/zookeeper-3.5.1-metrika.jar:/usr/share/zookeeper/slf4j-log4j12-1.7.5.jar:/usr/share/zookeeper/slf4j-api-1.7.5.jar:/usr/share/zookeeper/servlet-api-2.5-20081211.jar:/usr/share/zookeeper/netty-3.7.0.Final.jar:/usr/share/zookeeper/log4j-1.2.16.jar:/usr/share/zookeeper/jline-2.11.jar:/usr/share/zookeeper/jetty-util-6.1.26.jar:/usr/share/zookeeper/jetty-6.1.26.jar:/usr/share/zookeeper/javacc.jar:/usr/share/zookeeper/jackson-mapper-asl-1.9.11.jar:/usr/share/zookeeper/jackson-core-asl-1.9.11.jar:/usr/share/zookeeper/commons-cli-1.2.jar:/usr/src/java/lib/*.jar:/usr/etc/zookeeper"
-
-ZOOCFG="$ZOOCFGDIR/zoo.cfg"
-ZOO_LOG_DIR=/var/log/$NAME
-USER=zookeeper
-GROUP=zookeeper
-PIDDIR=/var/run/$NAME
-PIDFILE=$PIDDIR/$NAME.pid
-SCRIPTNAME=/etc/init.d/$NAME
-JAVA=/usr/bin/java
-ZOOMAIN="org.apache.zookeeper.server.quorum.QuorumPeerMain"
-ZOO_LOG4J_PROP="INFO,ROLLINGFILE"
-JMXLOCALONLY=false
-JAVA_OPTS="-Xms{{ '{{' }} cluster.get('xms','128M') {{ '{{' }} '}}' }} \
-    -Xmx{{ '{{' }} cluster.get('xmx','1G') {{ '{{' }} '}}' }} \
-    -Xloggc:/var/log/$NAME/zookeeper-gc.log \
-    -XX:+UseGCLogFileRotation \
-    -XX:NumberOfGCLogFiles=16 \
-    -XX:GCLogFileSize=16M \
-    -verbose:gc \
-    -XX:+PrintGCTimeStamps \
-    -XX:+PrintGCDateStamps \
-    -XX:+PrintGCDetails
-    -XX:+PrintTenuringDistribution \
-    -XX:+PrintGCApplicationStoppedTime \
-    -XX:+PrintGCApplicationConcurrentTime \
-    -XX:+PrintSafepointStatistics \
-    -XX:+UseParNewGC \
-    -XX:+UseConcMarkSweepGC \
--XX:+CMSParallelRemarkEnabled"
-```
-
-Salt init:
-
-``` text
-description "zookeeper-{{ '{{' }} cluster['name'] {{ '{{' }} '}}' }} centralized coordination service"
-
-start on runlevel [2345]
-stop on runlevel [!2345]
-
-respawn
-
-limit nofile 8192 8192
-
-pre-start script
-    [ -r "/etc/zookeeper-{{ '{{' }} cluster['name'] {{ '{{' }} '}}' }}/conf/environment" ] || exit 0
-    . /etc/zookeeper-{{ '{{' }} cluster['name'] {{ '{{' }} '}}' }}/conf/environment
-    [ -d $ZOO_LOG_DIR ] || mkdir -p $ZOO_LOG_DIR
-    chown $USER:$GROUP $ZOO_LOG_DIR
-end script
-
-script
-    . /etc/zookeeper-{{ '{{' }} cluster['name'] {{ '{{' }} '}}' }}/conf/environment
-    [ -r /etc/default/zookeeper ] && . /etc/default/zookeeper
-    if [ -z "$JMXDISABLE" ]; then
-        JAVA_OPTS="$JAVA_OPTS -Dcom.sun.management.jmxremote -Dcom.sun.management.jmxremote.local.only=$JMXLOCALONLY"
-    fi
-    exec start-stop-daemon --start -c $USER --exec $JAVA --name zookeeper-{{ '{{' }} cluster['name'] {{ '{{' }} '}}' }} \
-        -- -cp $CLASSPATH $JAVA_OPTS -Dzookeeper.log.dir=${ZOO_LOG_DIR} \
-        -Dzookeeper.root.logger=${ZOO_LOG4J_PROP} $ZOOMAIN $ZOOCFG
-end script
-```
-

docs/ru/operations/tips.md

@@ -0,0 +1 @@
+../../en/operations/tips.md

docs/ru/sql-reference/functions/array-functions.md

@@ -1528,3 +1528,52 @@ SELECT arrayAUC([0.1, 0.4, 0.35, 0.8], [0, 0, 1, 1]);
 └────────────────────────────────────────---──┘
 ```
 
+## arrayProduct {#arrayproduct}
+
+Возвращает произведение элементов [массива](../../sql-reference/data-types/array.md).
+
+**Синтаксис**
+
+``` sql
+arrayProduct(arr)
+```
+
+**Аргументы**
+
+-   `arr` — [массив](../../sql-reference/data-types/array.md) числовых значений.
+
+**Возвращаемое значение**
+
+-   Произведение элементов массива.
+
+Тип: [Float64](../../sql-reference/data-types/float.md).
+
+**Примеры**
+
+Запрос:
+
+``` sql
+SELECT arrayProduct([1,2,3,4,5,6]) as res;
+```
+
+Результат:
+
+``` text
+┌─res───┐
+│ 720   │
+└───────┘
+```
+
+Запрос:
+
+``` sql
+SELECT arrayProduct([toDecimal64(1,8), toDecimal64(2,8), toDecimal64(3,8)]) as res, toTypeName(res);
+```
+
+Возвращаемое значение всегда имеет тип [Float64](../../sql-reference/data-types/float.md). Результат:
+
+``` text
+┌─res─┬─toTypeName(arrayProduct(array(toDecimal64(1, 8), toDecimal64(2, 8), toDecimal64(3, 8))))─┐
+│ 6   │ Float64                                                                                  │
+└─────┴──────────────────────────────────────────────────────────────────────────────────────────┘
+```

docs/ru/sql-reference/statements/explain.md

@@ -47,9 +47,9 @@ Union
 
 ### EXPLAIN AST {#explain-ast}
 
-Дамп AST запроса.
+Дамп AST запроса. Поддерживает все типы запросов, не только `SELECT`.
 
-Пример:
+Примеры:
 
 ```sql
 EXPLAIN AST SELECT 1;
@@ -63,6 +63,22 @@ SelectWithUnionQuery (children 1)
     Literal UInt64_1
 ```
 
+```sql
+EXPLAIN AST ALTER TABLE t1 DELETE WHERE date = today();
+```
+
+```sql
+  explain
+  AlterQuery  t1 (children 1)
+   ExpressionList (children 1)
+    AlterCommand 27 (children 1)
+     Function equals (children 1)
+      ExpressionList (children 2)
+       Identifier date
+       Function today (children 1)
+        ExpressionList
+```
+
 ### EXPLAIN SYNTAX {#explain-syntax}
 
 Возвращает текст запроса после применения синтаксических оптимизаций.

programs/client/Client.cpp

@@ -1461,18 +1461,25 @@ private:
             // when `lambda()` function gets substituted into a wrong place.
             // To avoid dealing with these cases, run the check only for the
             // queries we were able to successfully execute.
-            // The final caveat is that sometimes WITH queries are not executed,
+            // Another caveat is that sometimes WITH queries are not executed,
             // if they are not referenced by the main SELECT, so they can still
             // have the aforementioned problems. Disable this check for such
             // queries, for lack of a better solution.
-            if (!have_error && queryHasWithClause(parsed_query.get()))
+            // There is also a problem that fuzzer substitutes positive Int64
+            // literals or Decimal literals, which are then parsed back as
+            // UInt64, and suddenly duplicate alias substitition starts or stops
+            // working (ASTWithAlias::formatImpl) or something like that.
+            // So we compare not even the first and second formatting of the
+            // query, but second and third.
+            // If you have to add any more workarounds to this check, just remove
+            // it altogether, it's not so useful.
+            if (!have_error && !queryHasWithClause(parsed_query.get()))
             {
-                ASTPtr parsed_formatted_query;
+                ASTPtr ast_2;
                 try
                 {
                     const auto * tmp_pos = query_to_send.c_str();
-                    parsed_formatted_query = parseQuery(tmp_pos,
-                        tmp_pos + query_to_send.size(),
+                    ast_2 = parseQuery(tmp_pos, tmp_pos + query_to_send.size(),
                         false /* allow_multi_statements */);
                 }
                 catch (Exception & e)
@@ -1483,25 +1490,31 @@ private:
                     }
                 }
 
-                if (parsed_formatted_query)
+                if (ast_2)
                 {
-                    const auto formatted_twice
-                        = parsed_formatted_query->formatForErrorMessage();
+                    const auto text_2 = ast_2->formatForErrorMessage();
+                    const auto * tmp_pos = text_2.c_str();
+                    const auto ast_3 = parseQuery(tmp_pos, tmp_pos + text_2.size(),
+                        false /* allow_multi_statements */);
+                    const auto text_3 = ast_3->formatForErrorMessage();
 
-                    if (formatted_twice != query_to_send)
+                    if (text_3 != text_2)
                     {
                         fmt::print(stderr, "The query formatting is broken.\n");
 
                         printChangedSettings();
 
-                        fmt::print(stderr, "Got the following (different) text after formatting the fuzzed query and parsing it back:\n'{}'\n, expected:\n'{}'\n",
-                            formatted_twice, query_to_send);
+                        fmt::print(stderr,
+                            "Got the following (different) text after formatting the fuzzed query and parsing it back:\n'{}'\n, expected:\n'{}'\n",
+                            text_3, text_2);
                         fmt::print(stderr, "In more detail:\n");
-                        fmt::print(stderr, "AST-1:\n'{}'\n", parsed_query->dumpTree());
+                        fmt::print(stderr, "AST-1 (generated by fuzzer):\n'{}'\n", parsed_query->dumpTree());
                         fmt::print(stderr, "Text-1 (AST-1 formatted):\n'{}'\n", query_to_send);
-                        fmt::print(stderr, "AST-2 (Text-1 parsed):\n'{}'\n", parsed_formatted_query->dumpTree());
-                        fmt::print(stderr, "Text-2 (AST-2 formatted):\n'{}'\n", formatted_twice);
-                        fmt::print(stderr, "Text-1 must be equal to Text-2, but it is not.\n");
+                        fmt::print(stderr, "AST-2 (Text-1 parsed):\n'{}'\n", ast_2->dumpTree());
+                        fmt::print(stderr, "Text-2 (AST-2 formatted):\n'{}'\n", text_2);
+                        fmt::print(stderr, "AST-3 (Text-2 parsed):\n'{}'\n", ast_3->dumpTree());
+                        fmt::print(stderr, "Text-3 (AST-3 formatted):\n'{}'\n", text_3);
+                        fmt::print(stderr, "Text-3 must be equal to Text-2, but it is not.\n");
 
                         exit(1);
                     }
@@ -1518,6 +1531,11 @@ private:
                 server_exception.reset();
                 client_exception.reset();
                 have_error = false;
+
+                // We have to reinitialize connection after errors, because it
+                // might have gotten into a wrong state and we'll get false
+                // positives about "Unknown packet from server".
+                connection->forceConnected(connection_parameters.timeouts);
             }
             else if (ast_to_process->formatForErrorMessage().size() > 500)
             {

programs/client/QueryFuzzer.h

@@ -4,11 +4,14 @@
 #include <unordered_map>
 #include <vector>
 
+#include <pcg-random/pcg_random.hpp>
+
 #include <Common/randomSeed.h>
 #include <Common/Stopwatch.h>
 #include <Core/Field.h>
 #include <Parsers/IAST.h>
 
+
 namespace DB
 {
 

programs/git-import/git-import.cpp

@@ -774,7 +774,7 @@ UInt128 diffHash(const CommitDiff & file_changes)
     }
 
     UInt128 hash_of_diff;
-    hasher.get128(hash_of_diff.low, hash_of_diff.high);
+    hasher.get128(hash_of_diff.items[0], hash_of_diff.items[1]);
 
     return hash_of_diff;
 }

programs/obfuscator/Obfuscator.cpp

@@ -365,16 +365,20 @@ static void transformFixedString(const UInt8 * src, UInt8 * dst, size_t size, UI
     }
 }
 
-static void transformUUID(const UInt128 & src, UInt128 & dst, UInt64 seed)
+static void transformUUID(const UUID & src_uuid, UUID & dst_uuid, UInt64 seed)
 {
+    const UInt128 & src = src_uuid.toUnderType();
+    UInt128 & dst = dst_uuid.toUnderType();
+
     SipHash hash;
     hash.update(seed);
-    hash.update(reinterpret_cast<const char *>(&src), sizeof(UInt128));
+    hash.update(reinterpret_cast<const char *>(&src), sizeof(UUID));
 
     /// Saving version and variant from an old UUID
     hash.get128(reinterpret_cast<char *>(&dst));
-    dst.high = (dst.high & 0x1fffffffffffffffull) | (src.high & 0xe000000000000000ull);
-    dst.low = (dst.low & 0xffffffffffff0fffull) | (src.low & 0x000000000000f000ull);
+
+    dst.items[1] = (dst.items[1] & 0x1fffffffffffffffull) | (src.items[1] & 0xe000000000000000ull);
+    dst.items[0] = (dst.items[0] & 0xffffffffffff0fffull) | (src.items[0] & 0x000000000000f000ull);
 }
 
 class FixedStringModel : public IModel
@@ -426,10 +430,10 @@ public:
 
     ColumnPtr generate(const IColumn & column) override
     {
-        const ColumnUInt128 & src_column = assert_cast<const ColumnUInt128 &>(column);
+        const ColumnUUID & src_column = assert_cast<const ColumnUUID &>(column);
         const auto & src_data = src_column.getData();
 
-        auto res_column = ColumnUInt128::create();
+        auto res_column = ColumnUUID::create();
         auto & res_data = res_column->getData();
 
         res_data.resize(src_data.size());

programs/odbc-bridge/ODBCBlockOutputStream.cpp

@@ -1,5 +1,6 @@
 #include "ODBCBlockOutputStream.h"
 
+#include <Common/hex.h>
 #include <common/logger_useful.h>
 #include <Core/Field.h>
 #include <common/LocalDate.h>
@@ -37,7 +38,6 @@ namespace
         query.IAST::format(settings);
         return buf.str();
     }
-
 }
 
 ODBCBlockOutputStream::ODBCBlockOutputStream(nanodbc::ConnectionHolderPtr connection_,

src/Access/IAccessStorage.h

@@ -1,7 +1,7 @@
 #pragma once
 
 #include <Access/IAccessEntity.h>
-#include <common/types.h>
+#include <Core/Types.h>
 #include <Core/UUID.h>
 #include <ext/scope_guard.h>
 #include <functional>

src/Access/RolesOrUsersSet.h

@@ -1,6 +1,7 @@
 #pragma once
 
 #include <Core/UUID.h>
+#include <Core/Types.h>
 #include <boost/container/flat_set.hpp>
 #include <memory>
 #include <optional>

src/Access/RowPolicy.h

@@ -2,6 +2,7 @@
 
 #include <Access/IAccessEntity.h>
 #include <Access/RolesOrUsersSet.h>
+#include <Core/Types.h>
 #include <array>
 
 

src/AggregateFunctions/AggregateFunctionAvg.h

@@ -7,7 +7,7 @@
 #include <DataTypes/DataTypesDecimal.h>
 #include <DataTypes/DataTypesNumber.h>
 #include <AggregateFunctions/IAggregateFunction.h>
-#include "Core/DecimalFunctions.h"
+#include <Core/DecimalFunctions.h>
 
 
 namespace DB

src/AggregateFunctions/AggregateFunctionAvgWeighted.h

@@ -5,18 +5,18 @@
 
 namespace DB
 {
-template <class T>
+template <typename T>
 using AvgWeightedFieldType = std::conditional_t<IsDecimalNumber<T>,
     std::conditional_t<std::is_same_v<T, Decimal256>, Decimal256, Decimal128>,
     std::conditional_t<DecimalOrExtendedInt<T>,
         Float64, // no way to do UInt128 * UInt128, better cast to Float64
         NearestFieldType<T>>>;
 
-template <class T, class U>
+template <typename T, typename U>
 using MaxFieldType = std::conditional_t<(sizeof(AvgWeightedFieldType<T>) > sizeof(AvgWeightedFieldType<U>)),
     AvgWeightedFieldType<T>, AvgWeightedFieldType<U>>;
 
-template <class Value, class Weight>
+template <typename Value, typename Weight>
 class AggregateFunctionAvgWeighted final :
     public AggregateFunctionAvgBase<
         MaxFieldType<Value, Weight>, AvgWeightedFieldType<Weight>, AggregateFunctionAvgWeighted<Value, Weight>>

src/AggregateFunctions/AggregateFunctionBoundingRatio.h

@@ -2,7 +2,6 @@
 
 #include <DataTypes/DataTypesNumber.h>
 #include <Columns/ColumnsNumber.h>
-#include <Common/FieldVisitors.h>
 #include <IO/ReadHelpers.h>
 #include <IO/WriteHelpers.h>
 #include <AggregateFunctions/Helpers.h>

src/AggregateFunctions/AggregateFunctionDistinct.h

@@ -10,7 +10,6 @@
 #include <Common/HashTable/HashMap.h>
 #include <Common/SipHash.h>
 
-#include <Common/FieldVisitors.h>
 
 namespace DB
 {

src/AggregateFunctions/AggregateFunctionGroupArray.h

@@ -296,12 +296,7 @@ public:
         if (size)
         {
             typename ColumnVector<T>::Container & data_to = assert_cast<ColumnVector<T> &>(arr_to.getData()).getData();
-            if constexpr (is_big_int_v<T>)
-                // is data_to empty? we should probably use std::vector::insert then
-                for (auto it = this->data(place).value.begin(); it != this->data(place).value.end(); it++)
-                    data_to.push_back(*it);
-            else
-                data_to.insert(this->data(place).value.begin(), this->data(place).value.end());
+            data_to.insert(this->data(place).value.begin(), this->data(place).value.end());
         }
     }
 

src/AggregateFunctions/AggregateFunctionGroupArrayMoving.cpp

@@ -21,6 +21,7 @@ namespace ErrorCodes
 namespace
 {
 
+/// TODO Proper support for Decimal256.
 template <typename T, typename LimitNumberOfElements>
 struct MovingSum
 {

src/AggregateFunctions/AggregateFunctionGroupArrayMoving.h

@@ -38,7 +38,7 @@ struct MovingData
     using Array = PODArray<T, 32, Allocator>;
 
     Array value;    /// Prefix sums.
-    T sum = 0;
+    T sum{};
 
     void NO_SANITIZE_UNDEFINED add(T val, Arena * arena)
     {
@@ -69,9 +69,9 @@ struct MovingAvgData : public MovingData<T>
     T NO_SANITIZE_UNDEFINED get(size_t idx, UInt64 window_size) const
     {
         if (idx < window_size)
-            return this->value[idx] / window_size;
+            return this->value[idx] / T(window_size);
         else
-            return (this->value[idx] - this->value[idx - window_size]) / window_size;
+            return (this->value[idx] - this->value[idx - window_size]) / T(window_size);
     }
 };
 

src/AggregateFunctions/AggregateFunctionMannWhitney.h

@@ -6,7 +6,6 @@
 #include <Columns/ColumnVector.h>
 #include <Columns/ColumnTuple.h>
 #include <Common/assert_cast.h>
-#include <Common/FieldVisitors.h>
 #include <Common/PODArray_fwd.h>
 #include <common/types.h>
 #include <DataTypes/DataTypesDecimal.h>
@@ -21,7 +20,7 @@
 
 #include <Common/ArenaAllocator.h>
 
-#include <iostream>
+
 namespace DB
 {
 

src/AggregateFunctions/AggregateFunctionQuantile.cpp

@@ -100,13 +100,14 @@ AggregateFunctionPtr createAggregateFunctionQuantile(const std::string & name, c
         if (which.idx == TypeIndex::Decimal32) return std::make_shared<Function<Decimal32, false>>(argument_types, params);
         if (which.idx == TypeIndex::Decimal64) return std::make_shared<Function<Decimal64, false>>(argument_types, params);
         if (which.idx == TypeIndex::Decimal128) return std::make_shared<Function<Decimal128, false>>(argument_types, params);
+        if (which.idx == TypeIndex::Decimal256) return std::make_shared<Function<Decimal256, false>>(argument_types, params);
         if (which.idx == TypeIndex::DateTime64) return std::make_shared<Function<DateTime64, false>>(argument_types, params);
-        //if (which.idx == TypeIndex::Decimal256) return std::make_shared<Function<Decimal256, false>>(argument_types, params);
     }
 
     if constexpr (supportBigInt<Function>())
     {
         if (which.idx == TypeIndex::Int128) return std::make_shared<Function<Int128, true>>(argument_types, params);
+        if (which.idx == TypeIndex::UInt128) return std::make_shared<Function<Int128, true>>(argument_types, params);
         if (which.idx == TypeIndex::Int256) return std::make_shared<Function<Int256, true>>(argument_types, params);
         if (which.idx == TypeIndex::UInt256) return std::make_shared<Function<UInt256, true>>(argument_types, params);
     }

src/AggregateFunctions/AggregateFunctionStatistics.cpp

@@ -36,7 +36,7 @@ AggregateFunctionPtr createAggregateFunctionStatisticsBinary(const std::string &
     assertNoParameters(name, parameters);
     assertBinary(name, argument_types);
 
-    AggregateFunctionPtr res(createWithTwoNumericTypes<FunctionTemplate>(*argument_types[0], *argument_types[1], argument_types));
+    AggregateFunctionPtr res(createWithTwoBasicNumericTypes<FunctionTemplate>(*argument_types[0], *argument_types[1], argument_types));
     if (!res)
         throw Exception("Illegal types " + argument_types[0]->getName() + " and " + argument_types[1]->getName()
             + " of arguments for aggregate function " + name, ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT);

src/AggregateFunctions/AggregateFunctionStatisticsSimple.cpp

@@ -41,7 +41,7 @@ AggregateFunctionPtr createAggregateFunctionStatisticsBinary(const std::string &
     assertNoParameters(name, parameters);
     assertBinary(name, argument_types);
 
-    AggregateFunctionPtr res(createWithTwoNumericTypes<FunctionTemplate>(*argument_types[0], *argument_types[1], argument_types));
+    AggregateFunctionPtr res(createWithTwoBasicNumericTypes<FunctionTemplate>(*argument_types[0], *argument_types[1], argument_types));
     if (!res)
         throw Exception("Illegal types " + argument_types[0]->getName() + " and " + argument_types[1]->getName()
             + " of arguments for aggregate function " + name, ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT);

src/AggregateFunctions/AggregateFunctionStatisticsSimple.h

@@ -131,11 +131,10 @@ public:
                 static_cast<ResultType>(static_cast<const ColVecT2 &>(*columns[1]).getData()[row_num]));
         else
         {
-            if constexpr (std::is_same_v<T1, Decimal256>)
+            if constexpr (IsDecimalNumber<T1>)
             {
                 this->data(place).add(static_cast<ResultType>(
-                    static_cast<const ColVecT1 &>(*columns[0]).getData()[row_num].value
-                ));
+                    static_cast<const ColVecT1 &>(*columns[0]).getData()[row_num].value));
             }
             else
                 this->data(place).add(

src/AggregateFunctions/AggregateFunctionTopK.cpp

@@ -2,9 +2,10 @@
 #include <AggregateFunctions/AggregateFunctionTopK.h>
 #include <AggregateFunctions/Helpers.h>
 #include <AggregateFunctions/FactoryHelpers.h>
+#include <Common/FieldVisitors.h>
 #include <DataTypes/DataTypeDate.h>
 #include <DataTypes/DataTypeDateTime.h>
-#include "registerAggregateFunctions.h"
+
 
 #define TOP_K_MAX_SIZE 0xFFFFFF
 

src/AggregateFunctions/AggregateFunctionTopK.h

@@ -10,7 +10,6 @@
 #include <Columns/ColumnArray.h>
 
 #include <Common/SpaceSaving.h>
-#include <Common/FieldVisitors.h>
 #include <Common/assert_cast.h>
 
 #include <AggregateFunctions/IAggregateFunction.h>

src/AggregateFunctions/AggregateFunctionUniq.cpp

@@ -132,6 +132,12 @@ void registerAggregateFunctionsUniq(AggregateFunctionFactory & factory)
 
     factory.registerFunction("uniqExact",
         {createAggregateFunctionUniq<true, AggregateFunctionUniqExactData, AggregateFunctionUniqExactData<String>>, properties});
+
+#if USE_DATASKETCHES
+    factory.registerFunction("uniqThetaSketch",
+        {createAggregateFunctionUniq<AggregateFunctionUniqThetaSketchData, AggregateFunctionUniqThetaSketchData>, properties});
+#endif
+
 }
 
 }

src/AggregateFunctions/AggregateFunctionUniq.h

@@ -22,6 +22,7 @@
 
 #include <AggregateFunctions/UniquesHashSet.h>
 #include <AggregateFunctions/IAggregateFunction.h>
+#include <AggregateFunctions/ThetaSketchData.h>
 #include <AggregateFunctions/UniqVariadicHash.h>
 
 
@@ -69,7 +70,7 @@ struct AggregateFunctionUniqHLL12Data<String>
 };
 
 template <>
-struct AggregateFunctionUniqHLL12Data<UInt128>
+struct AggregateFunctionUniqHLL12Data<UUID>
 {
     using Set = HyperLogLogWithSmallSetOptimization<UInt64, 16, 12>;
     Set set;
@@ -124,6 +125,19 @@ struct AggregateFunctionUniqExactData<String>
 };
 
 
+/// uniqThetaSketch
+#if USE_DATASKETCHES
+
+struct AggregateFunctionUniqThetaSketchData
+{
+    using Set = ThetaSketchData<UInt64>;
+    Set set;
+
+    static String getName() { return "uniqThetaSketch"; }
+};
+
+#endif
+
 namespace detail
 {
 
@@ -133,16 +147,14 @@ template <typename T> struct AggregateFunctionUniqTraits
 {
     static UInt64 hash(T x)
     {
-        if constexpr (std::is_same_v<T, UInt128>)
-        {
-            return sipHash64(x);
-        }
-        else if constexpr (std::is_same_v<T, Float32> || std::is_same_v<T, Float64>)
+        if constexpr (std::is_same_v<T, Float32> || std::is_same_v<T, Float64>)
         {
             return ext::bit_cast<UInt64>(x);
         }
         else if constexpr (sizeof(T) <= sizeof(UInt64))
+        {
             return x;
+        }
         else
             return DefaultHash64<T>(x);
     }
@@ -184,11 +196,17 @@ struct OneAdder
                 UInt128 key;
                 SipHash hash;
                 hash.update(value.data, value.size);
-                hash.get128(key.low, key.high);
+                hash.get128(key);
 
                 data.set.insert(key);
             }
         }
+#if USE_DATASKETCHES
+        else if constexpr (std::is_same_v<Data, AggregateFunctionUniqThetaSketchData>)
+        {
+            data.set.insertOriginal(column.getDataAt(row_num));
+        }
+#endif
     }
 };
 

src/AggregateFunctions/AggregateFunctionUniqCombined.cpp

@@ -3,11 +3,13 @@
 #include <AggregateFunctions/AggregateFunctionFactory.h>
 #include <AggregateFunctions/Helpers.h>
 
+#include <Common/FieldVisitors.h>
+
 #include <DataTypes/DataTypeDate.h>
 #include <DataTypes/DataTypeDateTime.h>
 
 #include <functional>
-#include "registerAggregateFunctions.h"
+
 
 namespace DB
 {

src/AggregateFunctions/AggregateFunctionUniqCombined.h

@@ -1,7 +1,6 @@
 #pragma once
 
 #include <Common/CombinedCardinalityEstimator.h>
-#include <Common/FieldVisitors.h>
 #include <Common/SipHash.h>
 #include <Common/typeid_cast.h>
 #include <Common/assert_cast.h>

src/AggregateFunctions/AggregateFunctionUniqUpTo.cpp

@@ -1,11 +1,11 @@
 #include <AggregateFunctions/AggregateFunctionFactory.h>
 #include <AggregateFunctions/Helpers.h>
 #include <AggregateFunctions/AggregateFunctionUniqUpTo.h>
+#include <Common/FieldVisitors.h>
 #include <DataTypes/DataTypeDate.h>
 #include <DataTypes/DataTypeDateTime.h>
 #include <DataTypes/DataTypeString.h>
 #include <DataTypes/DataTypeFixedString.h>
-#include "registerAggregateFunctions.h"
 
 
 namespace DB

src/AggregateFunctions/AggregateFunctionUniqUpTo.h

@@ -1,6 +1,7 @@
 #pragma once
 
-#include <Common/FieldVisitors.h>
+#include <common/unaligned.h>
+
 #include <Common/typeid_cast.h>
 #include <Common/assert_cast.h>
 
@@ -35,7 +36,7 @@ namespace DB
   */
 
 template <typename T>
-struct __attribute__((__packed__)) AggregateFunctionUniqUpToData
+struct AggregateFunctionUniqUpToData
 {
 /** If count == threshold + 1 - this means that it is "overflowed" (values greater than threshold).
   * In this case (for example, after calling the merge function), the `data` array does not necessarily contain the initialized values
@@ -43,7 +44,17 @@ struct __attribute__((__packed__)) AggregateFunctionUniqUpToData
   *   then set count to `threshold + 1`, and values from another state are not copied.
   */
     UInt8 count = 0;
-    T data[0];
+    char data_ptr[0];
+
+    T load(size_t i) const
+    {
+        return unalignedLoad<T>(data_ptr + i * sizeof(T));
+    }
+
+    void store(size_t i, const T & x)
+    {
+        unalignedStore<T>(data_ptr + i * sizeof(T), x);
+    }
 
     size_t size() const
     {
@@ -60,12 +71,12 @@ struct __attribute__((__packed__)) AggregateFunctionUniqUpToData
 
         /// Linear search for the matching element.
         for (size_t i = 0; i < count; ++i)
-            if (data[i] == x)
+            if (load(i) == x)
                 return;
 
         /// Did not find the matching element. If there is room for one more element, insert it.
         if (count < threshold)
-            data[count] = x;
+            store(count, x);
 
         /// After increasing count, the state may be overflowed.
         ++count;
@@ -84,7 +95,7 @@ struct __attribute__((__packed__)) AggregateFunctionUniqUpToData
         }
 
         for (size_t i = 0; i < rhs.count; ++i)
-            insert(rhs.data[i], threshold);
+            insert(rhs.load(i), threshold);
     }
 
     void write(WriteBuffer & wb, UInt8 threshold) const
@@ -93,7 +104,7 @@ struct __attribute__((__packed__)) AggregateFunctionUniqUpToData
 
         /// Write values only if the state is not overflowed. Otherwise, they are not needed, and only the fact that the state is overflowed is important.
         if (count <= threshold)
-            wb.write(reinterpret_cast<const char *>(data), count * sizeof(data[0]));
+            wb.write(data_ptr, count * sizeof(T));
     }
 
     void read(ReadBuffer & rb, UInt8 threshold)
@@ -101,7 +112,7 @@ struct __attribute__((__packed__)) AggregateFunctionUniqUpToData
         readBinary(count, rb);
 
         if (count <= threshold)
-            rb.read(reinterpret_cast<char *>(data), count * sizeof(data[0]));
+            rb.read(data_ptr, count * sizeof(T));
     }
 
     /// ALWAYS_INLINE is required to have better code layout for uniqUpTo function

src/AggregateFunctions/Helpers.h

@@ -15,12 +15,12 @@
     M(Float32) \
     M(Float64)
 
-// No UInt128 here because of the name conflict
 #define FOR_NUMERIC_TYPES(M) \
     M(UInt8) \
     M(UInt16) \
     M(UInt32) \
     M(UInt64) \
+    M(UInt128) \
     M(UInt256) \
     M(Int8) \
     M(Int16) \
@@ -109,6 +109,8 @@ static IAggregateFunction * createWithUnsignedIntegerType(const IDataType & argu
     if (which.idx == TypeIndex::UInt16) return new AggregateFunctionTemplate<UInt16, Data<UInt16>>(std::forward<TArgs>(args)...);
     if (which.idx == TypeIndex::UInt32) return new AggregateFunctionTemplate<UInt32, Data<UInt32>>(std::forward<TArgs>(args)...);
     if (which.idx == TypeIndex::UInt64) return new AggregateFunctionTemplate<UInt64, Data<UInt64>>(std::forward<TArgs>(args)...);
+    if (which.idx == TypeIndex::UInt128) return new AggregateFunctionTemplate<UInt128, Data<UInt128>>(std::forward<TArgs>(args)...);
+    if (which.idx == TypeIndex::UInt256) return new AggregateFunctionTemplate<UInt256, Data<UInt256>>(std::forward<TArgs>(args)...);
     return nullptr;
 }
 
@@ -119,11 +121,11 @@ static IAggregateFunction * createWithNumericBasedType(const IDataType & argumen
     if (f)
         return f;
 
-    /// expects that DataTypeDate based on UInt16, DataTypeDateTime based on UInt32 and UUID based on UInt128
+    /// expects that DataTypeDate based on UInt16, DataTypeDateTime based on UInt32
     WhichDataType which(argument_type);
     if (which.idx == TypeIndex::Date) return new AggregateFunctionTemplate<UInt16>(std::forward<TArgs>(args)...);
     if (which.idx == TypeIndex::DateTime) return new AggregateFunctionTemplate<UInt32>(std::forward<TArgs>(args)...);
-    if (which.idx == TypeIndex::UUID) return new AggregateFunctionTemplate<UInt128>(std::forward<TArgs>(args)...);
+    if (which.idx == TypeIndex::UUID) return new AggregateFunctionTemplate<UUID>(std::forward<TArgs>(args)...);
     return nullptr;
 }
 
@@ -184,6 +186,29 @@ static IAggregateFunction * createWithTwoNumericTypes(const IDataType & first_ty
     return nullptr;
 }
 
+template <typename FirstType, template <typename, typename> class AggregateFunctionTemplate, typename... TArgs>
+static IAggregateFunction * createWithTwoBasicNumericTypesSecond(const IDataType & second_type, TArgs && ... args)
+{
+    WhichDataType which(second_type);
+#define DISPATCH(TYPE) \
+    if (which.idx == TypeIndex::TYPE) return new AggregateFunctionTemplate<FirstType, TYPE>(std::forward<TArgs>(args)...);
+    FOR_BASIC_NUMERIC_TYPES(DISPATCH)
+#undef DISPATCH
+    return nullptr;
+}
+
+template <template <typename, typename> class AggregateFunctionTemplate, typename... TArgs>
+static IAggregateFunction * createWithTwoBasicNumericTypes(const IDataType & first_type, const IDataType & second_type, TArgs && ... args)
+{
+    WhichDataType which(first_type);
+#define DISPATCH(TYPE) \
+    if (which.idx == TypeIndex::TYPE) \
+        return createWithTwoBasicNumericTypesSecond<TYPE, AggregateFunctionTemplate>(second_type, std::forward<TArgs>(args)...);
+    FOR_BASIC_NUMERIC_TYPES(DISPATCH)
+#undef DISPATCH
+    return nullptr;
+}
+
 template <typename FirstType, template <typename, typename> class AggregateFunctionTemplate, typename... TArgs>
 static IAggregateFunction * createWithTwoNumericOrDateTypesSecond(const IDataType & second_type, TArgs && ... args)
 {
@@ -195,7 +220,7 @@ static IAggregateFunction * createWithTwoNumericOrDateTypesSecond(const IDataTyp
     if (which.idx == TypeIndex::Enum8) return new AggregateFunctionTemplate<FirstType, Int8>(std::forward<TArgs>(args)...);
     if (which.idx == TypeIndex::Enum16) return new AggregateFunctionTemplate<FirstType, Int16>(std::forward<TArgs>(args)...);
 
-    /// expects that DataTypeDate based on UInt16, DataTypeDateTime based on UInt32 and UUID based on UInt128
+    /// expects that DataTypeDate based on UInt16, DataTypeDateTime based on UInt32
     if (which.idx == TypeIndex::Date) return new AggregateFunctionTemplate<FirstType, UInt16>(std::forward<TArgs>(args)...);
     if (which.idx == TypeIndex::DateTime) return new AggregateFunctionTemplate<FirstType, UInt32>(std::forward<TArgs>(args)...);
 
@@ -216,7 +241,7 @@ static IAggregateFunction * createWithTwoNumericOrDateTypes(const IDataType & fi
     if (which.idx == TypeIndex::Enum16)
         return createWithTwoNumericOrDateTypesSecond<Int16, AggregateFunctionTemplate>(second_type, std::forward<TArgs>(args)...);
 
-    /// expects that DataTypeDate based on UInt16, DataTypeDateTime based on UInt32 and UUID based on UInt128
+    /// expects that DataTypeDate based on UInt16, DataTypeDateTime based on UInt32
     if (which.idx == TypeIndex::Date)
         return createWithTwoNumericOrDateTypesSecond<UInt16, AggregateFunctionTemplate>(second_type, std::forward<TArgs>(args)...);
     if (which.idx == TypeIndex::DateTime)

src/AggregateFunctions/QuantileReservoirSampler.h

@@ -54,7 +54,7 @@ struct QuantileReservoirSampler
     /// Get the value of the `level` quantile. The level must be between 0 and 1.
     Value get(Float64 level)
     {
-        return data.quantileInterpolated(level);
+        return Value(data.quantileInterpolated(level));
     }
 
     /// Get the `size` values of `levels` quantiles. Write `size` results starting with `result` address.
@@ -62,7 +62,7 @@ struct QuantileReservoirSampler
     void getMany(const Float64 * levels, const size_t * indices, size_t size, Value * result)
     {
         for (size_t i = 0; i < size; ++i)
-            result[indices[i]] = data.quantileInterpolated(levels[indices[i]]);
+            result[indices[i]] = Value(data.quantileInterpolated(levels[indices[i]]));
     }
 
     /// The same, but in the case of an empty state, NaN is returned.

src/AggregateFunctions/ReservoirSampler.h

@@ -131,12 +131,20 @@ public:
         size_t left_index = static_cast<size_t>(index);
         size_t right_index = left_index + 1;
         if (right_index == samples.size())
-            return static_cast<double>(samples[left_index]);
+        {
+            if constexpr (DB::IsDecimalNumber<T>)
+                return static_cast<double>(samples[left_index].value);
+            else
+                return static_cast<double>(samples[left_index]);
+        }
 
         double left_coef = right_index - index;
         double right_coef = index - left_index;
 
-        return static_cast<double>(samples[left_index]) * left_coef + static_cast<double>(samples[right_index]) * right_coef;
+        if constexpr (DB::IsDecimalNumber<T>)
+            return static_cast<double>(samples[left_index].value) * left_coef + static_cast<double>(samples[right_index].value) * right_coef;
+        else
+            return static_cast<double>(samples[left_index]) * left_coef + static_cast<double>(samples[right_index]) * right_coef;
     }
 
     void merge(const ReservoirSampler<T, OnEmpty> & b)

src/AggregateFunctions/ThetaSketchData.h

@@ -0,0 +1,119 @@
+#pragma once
+
+#if !defined(ARCADIA_BUILD)
+#    include <Common/config.h>
+#endif
+
+#if USE_DATASKETCHES
+
+#include <boost/noncopyable.hpp>
+#include <memory>
+#include <theta_sketch.hpp> // Y_IGNORE
+#include <theta_union.hpp> // Y_IGNORE
+
+
+namespace DB
+{
+
+
+template <typename Key>
+class ThetaSketchData : private boost::noncopyable
+{
+private:
+    std::unique_ptr<datasketches::update_theta_sketch> sk_update;
+    std::unique_ptr<datasketches::theta_union> sk_union;
+
+    inline datasketches::update_theta_sketch * getSkUpdate()
+    {
+        if (!sk_update)
+            sk_update = std::make_unique<datasketches::update_theta_sketch>(datasketches::update_theta_sketch::builder().build());
+        return sk_update.get();
+    }
+
+    inline datasketches::theta_union * getSkUnion()
+    {
+        if (!sk_union)
+            sk_union = std::make_unique<datasketches::theta_union>(datasketches::theta_union::builder().build());
+        return sk_union.get();
+    }
+
+public:
+    using value_type = Key;
+
+    ThetaSketchData() = default;
+    ~ThetaSketchData() = default;
+
+    /// Insert original value without hash, as `datasketches::update_theta_sketch.update` will do the hash internal.
+    void insertOriginal(const StringRef & value)
+    {
+        getSkUpdate()->update(value.data, value.size);
+    }
+
+    /// Note that `datasketches::update_theta_sketch.update` will do the hash again.
+    void insert(Key value)
+    {
+        getSkUpdate()->update(value);
+    }
+
+    UInt64 size() const
+    {
+        if (sk_union)
+            return static_cast<UInt64>(sk_union->get_result().get_estimate());
+        else if (sk_update)
+            return static_cast<UInt64>(sk_update->get_estimate());
+        else
+            return 0;
+    }
+
+    void merge(const ThetaSketchData & rhs)
+    {
+        datasketches::theta_union * u = getSkUnion();
+
+        if (sk_update)
+        {
+            u->update(*sk_update);
+            sk_update.reset(nullptr);
+        }
+
+        if (rhs.sk_update)
+            u->update(*rhs.sk_update);
+        else if (rhs.sk_union)
+            u->update(rhs.sk_union->get_result());
+    }
+
+    /// You can only call for an empty object.
+    void read(DB::ReadBuffer & in)
+    {
+        datasketches::compact_theta_sketch::vector_bytes bytes;
+        readVectorBinary(bytes, in);
+        if (!bytes.empty())
+        {
+            auto sk = datasketches::compact_theta_sketch::deserialize(bytes.data(), bytes.size());
+            getSkUnion()->update(sk);
+        }
+    }
+
+    void write(DB::WriteBuffer & out) const
+    {
+        if (sk_update)
+        {
+            auto bytes = sk_update->compact().serialize();
+            writeVectorBinary(bytes, out);
+        }
+        else if (sk_union)
+        {
+            auto bytes = sk_union->get_result().serialize();
+            writeVectorBinary(bytes, out);
+        }
+        else
+        {
+            datasketches::compact_theta_sketch::vector_bytes bytes;
+            writeVectorBinary(bytes, out);
+        }
+    }
+};
+
+
+}
+
+#endif

src/AggregateFunctions/UniqVariadicHash.h

@@ -3,7 +3,6 @@
 #include <city.h>
 #include <Core/Defines.h>
 #include <Common/SipHash.h>
-#include <Common/UInt128.h>
 #include <Common/assert_cast.h>
 #include <Columns/ColumnTuple.h>
 
@@ -107,7 +106,7 @@ struct UniqVariadicHash<true, false>
         }
 
         UInt128 key;
-        hash.get128(key.low, key.high);
+        hash.get128(key);
         return key;
     }
 };
@@ -131,7 +130,7 @@ struct UniqVariadicHash<true, true>
         }
 
         UInt128 key;
-        hash.get128(key.low, key.high);
+        hash.get128(key);
         return key;
     }
 };

src/Columns/ColumnDecimal.h

@@ -85,8 +85,8 @@ private:
     {}
 
 public:
-    const char * getFamilyName() const override { return TypeName<T>::get(); }
-    TypeIndex getDataType() const override { return TypeId<T>::value; }
+    const char * getFamilyName() const override { return TypeName<T>; }
+    TypeIndex getDataType() const override { return TypeId<T>; }
 
     bool isNumeric() const override { return false; }
     bool canBeInsideNullable() const override { return true; }

src/Columns/ColumnUnique.h

@@ -665,7 +665,7 @@ UInt128 ColumnUnique<ColumnType>::IncrementalHash::getHash(const ColumnType & co
             column.updateHashWithValue(i, sip_hash);
 
         std::lock_guard lock(mutex);
-        sip_hash.get128(hash.low, hash.high);
+        sip_hash.get128(hash);
         cur_hash = hash;
         num_added_rows.store(column_size);
     }

src/Columns/ColumnVector.cpp

@@ -33,6 +33,7 @@ namespace ErrorCodes
     extern const int PARAMETER_OUT_OF_BOUND;
     extern const int SIZES_OF_COLUMNS_DOESNT_MATCH;
     extern const int LOGICAL_ERROR;
+    extern const int NOT_IMPLEMENTED;
 }
 
 template <typename T>
@@ -154,7 +155,7 @@ void ColumnVector<T>::getPermutation(bool reverse, size_t limit, int nan_directi
     else
     {
         /// A case for radix sort
-        if constexpr (is_arithmetic_v<T> && !std::is_same_v<T, UInt128>)
+        if constexpr (is_arithmetic_v<T> && !is_big_int_v<T>)
         {
             /// Thresholds on size. Lower threshold is arbitrary. Upper threshold is chosen by the type for histogram counters.
             if (s >= 256 && s <= std::numeric_limits<UInt32>::max())
@@ -292,28 +293,37 @@ MutableColumnPtr ColumnVector<T>::cloneResized(size_t size) const
         memcpy(new_col.data.data(), data.data(), count * sizeof(data[0]));
 
         if (size > count)
-            memset(static_cast<void *>(&new_col.data[count]), static_cast<int>(ValueType()), (size - count) * sizeof(ValueType));
+            memset(static_cast<void *>(&new_col.data[count]), 0, (size - count) * sizeof(ValueType));
     }
 
     return res;
 }
 
 template <typename T>
-UInt64 ColumnVector<T>::get64(size_t n) const
+UInt64 ColumnVector<T>::get64(size_t n [[maybe_unused]]) const
 {
-    return ext::bit_cast<UInt64>(data[n]);
+    if constexpr (is_arithmetic_v<T>)
+        return ext::bit_cast<UInt64>(data[n]);
+    else
+        throw Exception(ErrorCodes::NOT_IMPLEMENTED, "Cannot get the value of {} as UInt64", TypeName<T>);
 }
 
 template <typename T>
-inline Float64 ColumnVector<T>::getFloat64(size_t n) const
+inline Float64 ColumnVector<T>::getFloat64(size_t n [[maybe_unused]]) const
 {
-    return static_cast<Float64>(data[n]);
+    if constexpr (is_arithmetic_v<T>)
+        return static_cast<Float64>(data[n]);
+    else
+        throw Exception(ErrorCodes::NOT_IMPLEMENTED, "Cannot get the value of {} as Float64", TypeName<T>);
 }
 
 template <typename T>
-Float32 ColumnVector<T>::getFloat32(size_t n) const
+Float32 ColumnVector<T>::getFloat32(size_t n [[maybe_unused]]) const
 {
-    return static_cast<Float32>(data[n]);
+    if constexpr (is_arithmetic_v<T>)
+        return static_cast<Float32>(data[n]);
+    else
+        throw Exception(ErrorCodes::NOT_IMPLEMENTED, "Cannot get the value of {} as Float32", TypeName<T>);
 }
 
 template <typename T>
@@ -482,8 +492,6 @@ void ColumnVector<T>::gather(ColumnGathererStream & gatherer)
 template <typename T>
 void ColumnVector<T>::getExtremes(Field & min, Field & max) const
 {
-    using FastRefT = std::conditional_t<is_big_int_v<T>, const T &, const T>;
-
     size_t size = data.size();
 
     if (size == 0)
@@ -504,7 +512,7 @@ void ColumnVector<T>::getExtremes(Field & min, Field & max) const
     T cur_min = NaNOrZero<T>();
     T cur_max = NaNOrZero<T>();
 
-    for (FastRefT x : data)
+    for (const T & x : data)
     {
         if (isNaN(x))
             continue;
@@ -569,5 +577,6 @@ template class ColumnVector<Int128>;
 template class ColumnVector<Int256>;
 template class ColumnVector<Float32>;
 template class ColumnVector<Float64>;
+template class ColumnVector<UUID>;
 
 }

src/Columns/ColumnVector.h

@@ -12,6 +12,12 @@
 namespace DB
 {
 
+namespace ErrorCodes
+{
+    extern const int NOT_IMPLEMENTED;
+}
+
+
 /** Stuff for comparing numbers.
   * Integer values are compared as usual.
   * Floating-point numbers are compared this way that NaNs always end up at the end
@@ -90,6 +96,7 @@ struct FloatCompareHelper
 template <class U> struct CompareHelper<Float32, U> : public FloatCompareHelper<Float32> {};
 template <class U> struct CompareHelper<Float64, U> : public FloatCompareHelper<Float64> {};
 
+
 /** A template for columns that use a simple array to store.
  */
 template <typename T>
@@ -118,7 +125,7 @@ private:
     ColumnVector(std::initializer_list<T> il) : data{il} {}
 
 public:
-    bool isNumeric() const override { return IsNumber<T>; }
+    bool isNumeric() const override { return is_arithmetic_v<T>; }
 
     size_t size() const override
     {
@@ -221,8 +228,8 @@ public:
         data.reserve(n);
     }
 
-    const char * getFamilyName() const override { return TypeName<T>::get(); }
-    TypeIndex getDataType() const override { return TypeId<T>::value; }
+    const char * getFamilyName() const override { return TypeName<T>; }
+    TypeIndex getDataType() const override { return TypeId<T>; }
 
     MutableColumnPtr cloneResized(size_t size) const override;
 
@@ -245,18 +252,27 @@ public:
     /// Out of range conversion is permitted.
     UInt64 NO_SANITIZE_UNDEFINED getUInt(size_t n) const override
     {
-        return UInt64(data[n]);
+        if constexpr (is_arithmetic_v<T>)
+            return UInt64(data[n]);
+        else
+            throw Exception(ErrorCodes::NOT_IMPLEMENTED, "Cannot get the value of {} as UInt", TypeName<T>);
     }
 
     /// Out of range conversion is permitted.
     Int64 NO_SANITIZE_UNDEFINED getInt(size_t n) const override
     {
-        return Int64(data[n]);
+        if constexpr (is_arithmetic_v<T>)
+            return Int64(data[n]);
+        else
+            throw Exception(ErrorCodes::NOT_IMPLEMENTED, "Cannot get the value of {} as Int", TypeName<T>);
     }
 
     bool getBool(size_t n) const override
     {
-        return bool(data[n]);
+        if constexpr (is_arithmetic_v<T>)
+            return bool(data[n]);
+        else
+            throw Exception(ErrorCodes::NOT_IMPLEMENTED, "Cannot get the value of {} as bool", TypeName<T>);
     }
 
     void insert(const Field & x) override
@@ -370,5 +386,6 @@ extern template class ColumnVector<Int128>;
 extern template class ColumnVector<Int256>;
 extern template class ColumnVector<Float32>;
 extern template class ColumnVector<Float64>;
+extern template class ColumnVector<UUID>;
 
 }

src/Columns/ColumnsNumber.h

@@ -26,6 +26,6 @@ using ColumnInt256 = ColumnVector<Int256>;
 using ColumnFloat32 = ColumnVector<Float32>;
 using ColumnFloat64 = ColumnVector<Float64>;
 
-using ColumnUUID = ColumnVector<UInt128>;
+using ColumnUUID = ColumnVector<UUID>;
 
 }

src/Columns/IColumnUnique.h

@@ -1,7 +1,6 @@
 #pragma once
 #include <optional>
 #include <Columns/IColumn.h>
-#include <Common/UInt128.h>
 
 namespace DB
 {

src/Columns/tests/gtest_weak_hash_32.cpp

@@ -14,11 +14,11 @@
 #include <DataTypes/DataTypesNumber.h>
 
 #include <Common/WeakHash.h>
+#include <Common/hex.h>
 
 #include <unordered_map>
 #include <iostream>
 #include <sstream>
-#include <Common/hex.h>
 
 
 using namespace DB;
@@ -27,20 +27,11 @@ template <typename T>
 void checkColumn(
     const WeakHash32::Container & hash,
     const PaddedPODArray<T> & eq_class,
-    const std::function<std::string(size_t)> & print_function,
     size_t allowed_collisions = 0,
     size_t max_collisions_to_print = 10)
 {
     ASSERT_EQ(hash.size(), eq_class.size());
 
-    auto print_for_row = [&](size_t row)
-    {
-        std::string res = "row: " + std::to_string(row);
-        res += " hash: " + std::to_string(hash[row]);
-        res += " value: " + print_function(row);
-        return res;
-    };
-
     /// Check equal rows has equal hash.
     {
         std::unordered_map<T, UInt32> map;
@@ -55,11 +46,7 @@ void checkColumn(
             else
             {
                 if (it->second != hash[i])
-                {
-                    std::cout << "Different hashes for the same equivalent class (" << size_t(val) << "):\n";
-                    std::cout << print_for_row(it->first) << '\n';
-                    std::cout << print_for_row(i) << std::endl;
-                }
+                    std::cout << "Different hashes for the same equivalent class (" << toString(val) << ")\n";
 
                 ASSERT_EQ(it->second, hash[i]);
             }
@@ -88,7 +75,6 @@ void checkColumn(
                 if (num_collisions <= max_collisions_to_print)
                 {
                     collisions_str << "Collision:\n";
-                    collisions_str << print_for_row(i) << std::endl;
                 }
 
                 if (num_collisions > allowed_collisions)
@@ -117,7 +103,7 @@ TEST(WeakHash32, ColumnVectorU8)
     WeakHash32 hash(col->size());
     col->updateWeakHash32(hash);
 
-    checkColumn(hash.getData(), col->getData(), [&](size_t row) { return std::to_string(col->getElement(row)); });
+    checkColumn(hash.getData(), col->getData());
 }
 
 TEST(WeakHash32, ColumnVectorI8)
@@ -134,7 +120,7 @@ TEST(WeakHash32, ColumnVectorI8)
     WeakHash32 hash(col->size());
     col->updateWeakHash32(hash);
 
-    checkColumn(hash.getData(), col->getData(), [&](size_t row) { return std::to_string(col->getElement(row)); });
+    checkColumn(hash.getData(), col->getData());
 }
 
 TEST(WeakHash32, ColumnVectorU16)
@@ -151,7 +137,7 @@ TEST(WeakHash32, ColumnVectorU16)
     WeakHash32 hash(col->size());
     col->updateWeakHash32(hash);
 
-    checkColumn(hash.getData(), col->getData(), [&](size_t row) { return std::to_string(col->getElement(row)); });
+    checkColumn(hash.getData(), col->getData());
 }
 
 TEST(WeakHash32, ColumnVectorI16)
@@ -168,7 +154,7 @@ TEST(WeakHash32, ColumnVectorI16)
     WeakHash32 hash(col->size());
     col->updateWeakHash32(hash);
 
-    checkColumn(hash.getData(), col->getData(), [&](size_t row) { return std::to_string(col->getElement(row)); });
+    checkColumn(hash.getData(), col->getData());
 }
 
 TEST(WeakHash32, ColumnVectorU32)
@@ -185,7 +171,7 @@ TEST(WeakHash32, ColumnVectorU32)
     WeakHash32 hash(col->size());
     col->updateWeakHash32(hash);
 
-    checkColumn(hash.getData(), col->getData(), [&](size_t row) { return std::to_string(col->getElement(row)); });
+    checkColumn(hash.getData(), col->getData());
 }
 
 TEST(WeakHash32, ColumnVectorI32)
@@ -202,7 +188,7 @@ TEST(WeakHash32, ColumnVectorI32)
     WeakHash32 hash(col->size());
     col->updateWeakHash32(hash);
 
-    checkColumn(hash.getData(), col->getData(), [&](size_t row) { return std::to_string(col->getElement(row)); });
+    checkColumn(hash.getData(), col->getData());
 }
 
 TEST(WeakHash32, ColumnVectorU64)
@@ -219,7 +205,7 @@ TEST(WeakHash32, ColumnVectorU64)
     WeakHash32 hash(col->size());
     col->updateWeakHash32(hash);
 
-    checkColumn(hash.getData(), col->getData(), [&](size_t row) { return std::to_string(col->getElement(row)); });
+    checkColumn(hash.getData(), col->getData());
 }
 
 TEST(WeakHash32, ColumnVectorI64)
@@ -236,7 +222,7 @@ TEST(WeakHash32, ColumnVectorI64)
     WeakHash32 hash(col->size());
     col->updateWeakHash32(hash);
 
-    checkColumn(hash.getData(), col->getData(), [&](size_t row) { return std::to_string(col->getElement(row)); });
+    checkColumn(hash.getData(), col->getData());
 }
 
 TEST(WeakHash32, ColumnVectorU128)
@@ -250,7 +236,9 @@ TEST(WeakHash32, ColumnVectorU128)
     {
         for (uint64_t i = 0; i < 65536; ++i)
         {
-            UInt128 val(i << 32u, i << 32u);
+            UInt128 val;
+            val.items[0] = i << 32u;
+            val.items[1] = i << 32u;
             data.push_back(val);
             eq_data.push_back(i);
         }
@@ -259,7 +247,7 @@ TEST(WeakHash32, ColumnVectorU128)
     WeakHash32 hash(col->size());
     col->updateWeakHash32(hash);
 
-    checkColumn(hash.getData(), eq_data, [&](size_t row) { return col->getElement(row).toHexString(); });
+    checkColumn(hash.getData(), eq_data);
 }
 
 TEST(WeakHash32, ColumnVectorI128)
@@ -276,7 +264,7 @@ TEST(WeakHash32, ColumnVectorI128)
     WeakHash32 hash(col->size());
     col->updateWeakHash32(hash);
 
-    checkColumn(hash.getData(), col->getData(), [&](size_t row) { return std::to_string(Int64(col->getElement(row))); });
+    checkColumn(hash.getData(), col->getData());
 }
 
 TEST(WeakHash32, ColumnDecimal32)
@@ -293,7 +281,7 @@ TEST(WeakHash32, ColumnDecimal32)
     WeakHash32 hash(col->size());
     col->updateWeakHash32(hash);
 
-    checkColumn(hash.getData(), col->getData(), [&](size_t row) { return std::to_string(col->getElement(row)); });
+    checkColumn(hash.getData(), col->getData());
 }
 
 TEST(WeakHash32, ColumnDecimal64)
@@ -310,7 +298,7 @@ TEST(WeakHash32, ColumnDecimal64)
     WeakHash32 hash(col->size());
     col->updateWeakHash32(hash);
 
-    checkColumn(hash.getData(), col->getData(), [&](size_t row) { return std::to_string(col->getElement(row)); });
+    checkColumn(hash.getData(), col->getData());
 }
 
 TEST(WeakHash32, ColumnDecimal128)
@@ -327,7 +315,7 @@ TEST(WeakHash32, ColumnDecimal128)
     WeakHash32 hash(col->size());
     col->updateWeakHash32(hash);
 
-    checkColumn(hash.getData(), col->getData(), [&](size_t row) { return std::to_string(Int64(col->getElement(row))); });
+    checkColumn(hash.getData(), col->getData());
 }
 
 TEST(WeakHash32, ColumnString1)
@@ -349,7 +337,7 @@ TEST(WeakHash32, ColumnString1)
     WeakHash32 hash(col->size());
     col->updateWeakHash32(hash);
 
-    checkColumn(hash.getData(), data, [&](size_t row) { return col->getDataAt(row).toString(); });
+    checkColumn(hash.getData(), data);
 }
 
 TEST(WeakHash32, ColumnString2)
@@ -389,7 +377,7 @@ TEST(WeakHash32, ColumnString2)
     /// Now there is single collision between 'k' * 544 and 'q' * 2512 (which is calculated twice)
     size_t allowed_collisions = 4;
 
-    checkColumn(hash.getData(), data, [&](size_t row) { return col->getDataAt(row).toString(); }, allowed_collisions);
+    checkColumn(hash.getData(), data, allowed_collisions);
 }
 
 TEST(WeakHash32, ColumnString3)
@@ -427,7 +415,7 @@ TEST(WeakHash32, ColumnString3)
     WeakHash32 hash(col->size());
     col->updateWeakHash32(hash);
 
-    checkColumn(hash.getData(), data, [&](size_t row) { return col->getDataAt(row).toString(); });
+    checkColumn(hash.getData(), data);
 }
 
 TEST(WeakHash32, ColumnFixedString)
@@ -455,7 +443,7 @@ TEST(WeakHash32, ColumnFixedString)
     WeakHash32 hash(col->size());
     col->updateWeakHash32(hash);
 
-    checkColumn(hash.getData(), data, [&](size_t row) { return col->getDataAt(row).toString(); });
+    checkColumn(hash.getData(), data);
 }
 
 TEST(WeakHash32, ColumnArray)
@@ -502,15 +490,7 @@ TEST(WeakHash32, ColumnArray)
     WeakHash32 hash(col_arr->size());
     col_arr->updateWeakHash32(hash);
 
-    auto print_function = [&col_arr](size_t row)
-    {
-        auto & offsets = col_arr->getOffsets();
-        size_t s = offsets[row] - offsets[row - 1];
-        auto value = col_arr->getData().getUInt(offsets[row]);
-        return std::string("[array of size ") + std::to_string(s) + " with values " + std::to_string(value) + "]";
-    };
-
-    checkColumn(hash.getData(), eq_data, print_function);
+    checkColumn(hash.getData(), eq_data);
 }
 
 TEST(WeakHash32, ColumnArray2)
@@ -545,15 +525,7 @@ TEST(WeakHash32, ColumnArray2)
     WeakHash32 hash(col_arr->size());
     col_arr->updateWeakHash32(hash);
 
-    auto print_function = [&col_arr](size_t row)
-    {
-        auto & offsets = col_arr->getOffsets();
-        auto value1 = col_arr->getData().getUInt(offsets[row]);
-        auto value2 = col_arr->getData().getUInt(offsets[row] + 1);
-        return std::string("[") + std::to_string(value1) + ", " + std::to_string(value2) + "]";
-    };
-
-    checkColumn(hash.getData(), eq_data, print_function);
+    checkColumn(hash.getData(), eq_data);
 }
 
 TEST(WeakHash32, ColumnArrayArray)
@@ -610,20 +582,7 @@ TEST(WeakHash32, ColumnArrayArray)
     WeakHash32 hash(col_arr_arr->size());
     col_arr_arr->updateWeakHash32(hash);
 
-    auto print_function = [&col_arr_arr](size_t row2)
-    {
-        auto & offsets2 = col_arr_arr->getOffsets();
-        size_t s2 = offsets2[row2] - offsets2[row2 - 1];
-        const auto & arr2 = typeid_cast<const ColumnArray &>(col_arr_arr->getData());
-        const auto & offsets = arr2.getOffsets();
-        size_t row = offsets2[row2];
-        size_t s = offsets[row] - offsets[row - 1];
-        auto value = arr2.getData().getUInt(offsets[row]);
-        return std::string("[array of size ") + std::to_string(s2) + " with values ["
-                                "[[array of size " + std::to_string(s) + " with values " + std::to_string(value) + "]]";
-    };
-
-    checkColumn(hash.getData(), eq_data, print_function);
+    checkColumn(hash.getData(), eq_data);
 }
 
 TEST(WeakHash32, ColumnConst)
@@ -642,7 +601,7 @@ TEST(WeakHash32, ColumnConst)
     WeakHash32 hash(col_const->size());
     col_const->updateWeakHash32(hash);
 
-    checkColumn(hash.getData(), data, [&](size_t) { return std::to_string(0); });
+    checkColumn(hash.getData(), data);
 }
 
 TEST(WeakHash32, ColumnLowcardinality)
@@ -663,7 +622,7 @@ TEST(WeakHash32, ColumnLowcardinality)
     WeakHash32 hash(col->size());
     col->updateWeakHash32(hash);
 
-    checkColumn(hash.getData(), data, [&](size_t row) { return std::to_string(col->getUInt(row)); });
+    checkColumn(hash.getData(), data);
 }
 
 TEST(WeakHash32, ColumnNullable)
@@ -689,7 +648,7 @@ TEST(WeakHash32, ColumnNullable)
     WeakHash32 hash(col_null->size());
     col_null->updateWeakHash32(hash);
 
-    checkColumn(hash.getData(), eq, [&](size_t row) { return eq[row] == -1 ? "Null" : std::to_string(eq[row]); });
+    checkColumn(hash.getData(), eq);
 }
 
 TEST(WeakHash32, ColumnTupleUInt64UInt64)
@@ -720,14 +679,7 @@ TEST(WeakHash32, ColumnTupleUInt64UInt64)
     WeakHash32 hash(col_tuple->size());
     col_tuple->updateWeakHash32(hash);
 
-    auto print_func = [&](size_t row)
-    {
-        std::string l = std::to_string(col_tuple->getColumn(0).getUInt(row));
-        std::string r = std::to_string(col_tuple->getColumn(1).getUInt(row));
-        return "(" + l + ", " + r + ")";
-    };
-
-    checkColumn(hash.getData(), eq, print_func);
+    checkColumn(hash.getData(), eq);
 }
 
 TEST(WeakHash32, ColumnTupleUInt64String)
@@ -765,15 +717,8 @@ TEST(WeakHash32, ColumnTupleUInt64String)
     WeakHash32 hash(col_tuple->size());
     col_tuple->updateWeakHash32(hash);
 
-    auto print_func = [&](size_t row)
-    {
-        std::string l = std::to_string(col_tuple->getColumn(0).getUInt(row));
-        std::string r = col_tuple->getColumn(1).getDataAt(row).toString();
-        return "(" + l + ", " + r + ")";
-    };
-
     size_t allowed_collisions = 8;
-    checkColumn(hash.getData(), eq, print_func, allowed_collisions);
+    checkColumn(hash.getData(), eq, allowed_collisions);
 }
 
 TEST(WeakHash32, ColumnTupleUInt64FixedString)
@@ -811,14 +756,7 @@ TEST(WeakHash32, ColumnTupleUInt64FixedString)
     WeakHash32 hash(col_tuple->size());
     col_tuple->updateWeakHash32(hash);
 
-    auto print_func = [&](size_t row)
-    {
-        std::string l = std::to_string(col_tuple->getColumn(0).getUInt(row));
-        std::string r = col_tuple->getColumn(1).getDataAt(row).toString();
-        return "(" + l + ", " + r + ")";
-    };
-
-    checkColumn(hash.getData(), eq, print_func);
+    checkColumn(hash.getData(), eq);
 }
 
 TEST(WeakHash32, ColumnTupleUInt64Array)
@@ -865,23 +803,10 @@ TEST(WeakHash32, ColumnTupleUInt64Array)
     WeakHash32 hash(col_tuple->size());
     col_tuple->updateWeakHash32(hash);
 
-    auto print_func = [&](size_t row)
-    {
-        std::string l = std::to_string(col_tuple->getColumn(0).getUInt(row));
-
-        const auto * col_arr = typeid_cast<const ColumnArray *>(col_tuple->getColumnPtr(1).get());
-        const auto & offsets = col_arr->getOffsets();
-        size_t s = offsets[row] - offsets[row - 1];
-        auto value = col_arr->getData().getUInt(offsets[row]);
-        auto r = std::string("[array of size ") + std::to_string(s) + " with values " + std::to_string(value) + "]";
-
-        return "(" + l + ", " + r + ")";
-    };
-
     /// There are 2 collisions right now (repeated 2 times each):
     /// (0, [array of size 1212 with values 7]) vs (0, [array of size 2265 with values 17])
     /// (0, [array of size 558 with values 5]) vs (1, [array of size 879 with values 21])
 
     size_t allowed_collisions = 8;
-    checkColumn(hash.getData(), eq_data, print_func, allowed_collisions);
+    checkColumn(hash.getData(), eq_data, allowed_collisions);
 }

src/Common/ColumnsHashing.h

@@ -166,8 +166,7 @@ public:
         size_t operator()(const DictionaryKey & key) const
         {
             SipHash hash;
-            hash.update(key.hash.low);
-            hash.update(key.hash.high);
+            hash.update(key.hash);
             hash.update(key.size);
             return hash.get64();
         }

src/Common/FieldVisitors.cpp

@@ -1,4 +1,3 @@
-#include <Core/UUID.h>
 #include <IO/ReadBuffer.h>
 #include <IO/WriteBuffer.h>
 #include <IO/WriteHelpers.h>
@@ -27,7 +26,7 @@ template <typename T>
 static inline String formatQuotedWithPrefix(T x, const char * prefix)
 {
     WriteBufferFromOwnString wb;
-    wb.write(prefix, strlen(prefix));
+    writeCString(prefix, wb);
     writeQuoted(x, wb);
     return wb.str();
 }
@@ -48,10 +47,11 @@ String FieldVisitorDump::operator() (const DecimalField<Decimal32> & x) const {
 String FieldVisitorDump::operator() (const DecimalField<Decimal64> & x) const { return formatQuotedWithPrefix(x, "Decimal64_"); }
 String FieldVisitorDump::operator() (const DecimalField<Decimal128> & x) const { return formatQuotedWithPrefix(x, "Decimal128_"); }
 String FieldVisitorDump::operator() (const DecimalField<Decimal256> & x) const { return formatQuotedWithPrefix(x, "Decimal256_"); }
+String FieldVisitorDump::operator() (const UInt128 & x) const { return formatQuotedWithPrefix(x, "UInt128_"); }
 String FieldVisitorDump::operator() (const UInt256 & x) const { return formatQuotedWithPrefix(x, "UInt256_"); }
-String FieldVisitorDump::operator() (const Int256 & x) const { return formatQuotedWithPrefix(x, "Int256_"); }
 String FieldVisitorDump::operator() (const Int128 & x) const { return formatQuotedWithPrefix(x, "Int128_"); }
-String FieldVisitorDump::operator() (const UInt128 & x) const { return formatQuotedWithPrefix(UUID(x), "UUID_"); }
+String FieldVisitorDump::operator() (const Int256 & x) const { return formatQuotedWithPrefix(x, "Int256_"); }
+String FieldVisitorDump::operator() (const UUID & x) const { return formatQuotedWithPrefix(x, "UUID_"); }
 
 
 String FieldVisitorDump::operator() (const String & x) const
@@ -152,13 +152,11 @@ String FieldVisitorToString::operator() (const DecimalField<Decimal64> & x) cons
 String FieldVisitorToString::operator() (const DecimalField<Decimal128> & x) const { return formatQuoted(x); }
 String FieldVisitorToString::operator() (const DecimalField<Decimal256> & x) const { return formatQuoted(x); }
 String FieldVisitorToString::operator() (const Int128 & x) const { return formatQuoted(x); }
-String FieldVisitorToString::operator() (const UInt128 & x) const { return formatQuoted(UUID(x)); }
-String FieldVisitorToString::operator() (const AggregateFunctionStateData & x) const
-{
-    return formatQuoted(x.data);
-}
+String FieldVisitorToString::operator() (const UInt128 & x) const { return formatQuoted(x); }
 String FieldVisitorToString::operator() (const UInt256 & x) const { return formatQuoted(x); }
 String FieldVisitorToString::operator() (const Int256 & x) const { return formatQuoted(x); }
+String FieldVisitorToString::operator() (const UUID & x) const { return formatQuoted(x); }
+String FieldVisitorToString::operator() (const AggregateFunctionStateData & x) const { return formatQuoted(x.data); }
 
 String FieldVisitorToString::operator() (const Array & x) const
 {
@@ -228,6 +226,7 @@ void FieldVisitorWriteBinary::operator() (const UInt128 & x, WriteBuffer & buf)
 void FieldVisitorWriteBinary::operator() (const Int128 & x, WriteBuffer & buf) const { DB::writeVarInt(x, buf); }
 void FieldVisitorWriteBinary::operator() (const UInt256 & x, WriteBuffer & buf) const { DB::writeBinary(x, buf); }
 void FieldVisitorWriteBinary::operator() (const Int256 & x, WriteBuffer & buf) const { DB::writeBinary(x, buf); }
+void FieldVisitorWriteBinary::operator() (const UUID & x, WriteBuffer & buf) const { DB::writeBinary(x, buf); }
 void FieldVisitorWriteBinary::operator() (const DecimalField<Decimal32> & x, WriteBuffer & buf) const { DB::writeBinary(x.getValue(), buf); }
 void FieldVisitorWriteBinary::operator() (const DecimalField<Decimal64> & x, WriteBuffer & buf) const { DB::writeBinary(x.getValue(), buf); }
 void FieldVisitorWriteBinary::operator() (const DecimalField<Decimal128> & x, WriteBuffer & buf) const { DB::writeBinary(x.getValue(), buf); }
@@ -315,6 +314,13 @@ void FieldVisitorHash::operator() (const Int128 & x) const
     hash.update(x);
 }
 
+void FieldVisitorHash::operator() (const UUID & x) const
+{
+    UInt8 type = Field::Types::UUID;
+    hash.update(type);
+    hash.update(x);
+}
+
 void FieldVisitorHash::operator() (const Float64 & x) const
 {
     UInt8 type = Field::Types::Float64;

src/Common/FieldVisitors.h

@@ -19,14 +19,6 @@ namespace ErrorCodes
     extern const int NOT_IMPLEMENTED;
 }
 
-#pragma GCC diagnostic push
-#pragma GCC diagnostic ignored "-Wredundant-decls"
-// Just dont mess with it. If the redundant redeclaration is removed then ReaderHelpers.h should be included.
-// This leads to Arena.h inclusion which has a problem with ASAN stuff included properly and messing macro definition
-// which intefrers with... You dont want to know, really.
-UInt128 stringToUUID(const String & str);
-#pragma GCC diagnostic pop
-
 
 /** StaticVisitor (and its descendants) - class with overloaded operator() for all types of fields.
   * You could call visitor for field using function 'applyVisitor'.
@@ -71,8 +63,11 @@ public:
     String operator() (const Null & x) const;
     String operator() (const UInt64 & x) const;
     String operator() (const UInt128 & x) const;
+    String operator() (const UInt256 & x) const;
     String operator() (const Int64 & x) const;
     String operator() (const Int128 & x) const;
+    String operator() (const Int256 & x) const;
+    String operator() (const UUID & x) const;
     String operator() (const Float64 & x) const;
     String operator() (const String & x) const;
     String operator() (const Array & x) const;
@@ -83,9 +78,6 @@ public:
     String operator() (const DecimalField<Decimal128> & x) const;
     String operator() (const DecimalField<Decimal256> & x) const;
     String operator() (const AggregateFunctionStateData & x) const;
-
-    String operator() (const UInt256 & x) const;
-    String operator() (const Int256 & x) const;
 };
 
 
@@ -95,8 +87,11 @@ public:
     void operator() (const Null & x, WriteBuffer & buf) const;
     void operator() (const UInt64 & x, WriteBuffer & buf) const;
     void operator() (const UInt128 & x, WriteBuffer & buf) const;
+    void operator() (const UInt256 & x, WriteBuffer & buf) const;
     void operator() (const Int64 & x, WriteBuffer & buf) const;
     void operator() (const Int128 & x, WriteBuffer & buf) const;
+    void operator() (const Int256 & x, WriteBuffer & buf) const;
+    void operator() (const UUID & x, WriteBuffer & buf) const;
     void operator() (const Float64 & x, WriteBuffer & buf) const;
     void operator() (const String & x, WriteBuffer & buf) const;
     void operator() (const Array & x, WriteBuffer & buf) const;
@@ -107,9 +102,6 @@ public:
     void operator() (const DecimalField<Decimal128> & x, WriteBuffer & buf) const;
     void operator() (const DecimalField<Decimal256> & x, WriteBuffer & buf) const;
     void operator() (const AggregateFunctionStateData & x, WriteBuffer & buf) const;
-
-    void operator() (const UInt256 & x, WriteBuffer & buf) const;
-    void operator() (const Int256 & x, WriteBuffer & buf) const;
 };
 
 
@@ -120,8 +112,11 @@ public:
     String operator() (const Null & x) const;
     String operator() (const UInt64 & x) const;
     String operator() (const UInt128 & x) const;
+    String operator() (const UInt256 & x) const;
     String operator() (const Int64 & x) const;
     String operator() (const Int128 & x) const;
+    String operator() (const Int256 & x) const;
+    String operator() (const UUID & x) const;
     String operator() (const Float64 & x) const;
     String operator() (const String & x) const;
     String operator() (const Array & x) const;
@@ -132,9 +127,6 @@ public:
     String operator() (const DecimalField<Decimal128> & x) const;
     String operator() (const DecimalField<Decimal256> & x) const;
     String operator() (const AggregateFunctionStateData & x) const;
-
-    String operator() (const UInt256 & x) const;
-    String operator() (const Int256 & x) const;
 };
 
 
@@ -171,6 +163,7 @@ public:
     T operator() (const UInt64 & x) const { return T(x); }
     T operator() (const Int64 & x) const { return T(x); }
     T operator() (const Int128 & x) const { return T(x); }
+    T operator() (const UUID & x) const { return T(x.toUnderType()); }
 
     T operator() (const Float64 & x) const
     {
@@ -259,8 +252,11 @@ public:
     void operator() (const Null & x) const;
     void operator() (const UInt64 & x) const;
     void operator() (const UInt128 & x) const;
+    void operator() (const UInt256 & x) const;
     void operator() (const Int64 & x) const;
     void operator() (const Int128 & x) const;
+    void operator() (const Int256 & x) const;
+    void operator() (const UUID & x) const;
     void operator() (const Float64 & x) const;
     void operator() (const String & x) const;
     void operator() (const Array & x) const;
@@ -271,9 +267,6 @@ public:
     void operator() (const DecimalField<Decimal128> & x) const;
     void operator() (const DecimalField<Decimal256> & x) const;
     void operator() (const AggregateFunctionStateData & x) const;
-
-    void operator() (const UInt256 & x) const;
-    void operator() (const Int256 & x) const;
 };
 
 
@@ -309,7 +302,7 @@ public:
     bool operator() (Array &) const { throw Exception("Cannot sum Arrays", ErrorCodes::LOGICAL_ERROR); }
     bool operator() (Tuple &) const { throw Exception("Cannot sum Tuples", ErrorCodes::LOGICAL_ERROR); }
     bool operator() (Map &) const { throw Exception("Cannot sum Maps", ErrorCodes::LOGICAL_ERROR); }
-    bool operator() (UInt128 &) const { throw Exception("Cannot sum UUIDs", ErrorCodes::LOGICAL_ERROR); }
+    bool operator() (UUID &) const { throw Exception("Cannot sum UUIDs", ErrorCodes::LOGICAL_ERROR); }
     bool operator() (AggregateFunctionStateData &) const { throw Exception("Cannot sum AggregateFunctionStates", ErrorCodes::LOGICAL_ERROR); }
 
     bool operator() (Int128 & x) const

src/Common/FieldVisitorsAccurateComparison.h

@@ -36,41 +36,30 @@ public:
             if constexpr (std::is_arithmetic_v<T> && std::is_arithmetic_v<U>)
                 return accurate::equalsOp(l, r);
 
+            /// TODO This is wrong (does not respect scale).
             if constexpr (isDecimalField<T>() && isDecimalField<U>())
                 return l == r;
 
             if constexpr (isDecimalField<T>() && std::is_arithmetic_v<U>)
-                return l == DecimalField<Decimal128>(r, 0);
+                return l == DecimalField<Decimal256>(Decimal256(r), 0);
 
             if constexpr (std::is_arithmetic_v<T> && isDecimalField<U>())
-                return DecimalField<Decimal128>(l, 0) == r;
+                return DecimalField<Decimal256>(Decimal256(l), 0) == r;
 
-            if constexpr (std::is_same_v<T, String>)
+            if constexpr (std::is_same_v<T, String> && std::is_arithmetic_v<U>)
             {
-                if constexpr (std::is_same_v<U, UInt128>)
-                    return stringToUUID(l) == r;
-
-                if constexpr (std::is_arithmetic_v<U>)
-                {
-                    ReadBufferFromString in(l);
-                    U parsed;
-                    readText(parsed, in);
-                    return operator()(parsed, r);
-                }
+                ReadBufferFromString in(l);
+                U parsed;
+                readText(parsed, in);
+                return operator()(parsed, r);
             }
 
-            if constexpr (std::is_same_v<U, String>)
+            if constexpr (std::is_same_v<U, String> && std::is_arithmetic_v<T>)
             {
-                if constexpr (std::is_same_v<T, UInt128>)
-                    return l == stringToUUID(r);
-
-                if constexpr (std::is_arithmetic_v<T>)
-                {
-                    ReadBufferFromString in(r);
-                    T parsed;
-                    readText(parsed, in);
-                    return operator()(l, parsed);
-                }
+                ReadBufferFromString in(r);
+                T parsed;
+                readText(parsed, in);
+                return operator()(l, parsed);
             }
         }
 
@@ -96,41 +85,30 @@ public:
             if constexpr (std::is_arithmetic_v<T> && std::is_arithmetic_v<U>)
                 return accurate::lessOp(l, r);
 
+            /// TODO This is wrong (does not respect scale).
             if constexpr (isDecimalField<T>() && isDecimalField<U>())
                 return l < r;
 
             if constexpr (isDecimalField<T>() && std::is_arithmetic_v<U>)
-                return l < DecimalField<Decimal128>(r, 0);
+                return l < DecimalField<Decimal256>(Decimal256(r), 0);
 
             if constexpr (std::is_arithmetic_v<T> && isDecimalField<U>())
-                return DecimalField<Decimal128>(l, 0) < r;
+                return DecimalField<Decimal256>(Decimal256(l), 0) < r;
 
-            if constexpr (std::is_same_v<T, String>)
+            if constexpr (std::is_same_v<T, String> && std::is_arithmetic_v<U>)
             {
-                if constexpr (std::is_same_v<U, UInt128>)
-                    return stringToUUID(l) < r;
-
-                if constexpr (std::is_arithmetic_v<U>)
-                {
-                    ReadBufferFromString in(l);
-                    U parsed;
-                    readText(parsed, in);
-                    return operator()(parsed, r);
-                }
+                ReadBufferFromString in(l);
+                U parsed;
+                readText(parsed, in);
+                return operator()(parsed, r);
             }
 
-            if constexpr (std::is_same_v<U, String>)
+            if constexpr (std::is_same_v<U, String> && std::is_arithmetic_v<T>)
             {
-                if constexpr (std::is_same_v<T, UInt128>)
-                    return l < stringToUUID(r);
-
-                if constexpr (std::is_arithmetic_v<T>)
-                {
-                    ReadBufferFromString in(r);
-                    T parsed;
-                    readText(parsed, in);
-                    return operator()(l, parsed);
-                }
+                ReadBufferFromString in(r);
+                T parsed;
+                readText(parsed, in);
+                return operator()(l, parsed);
             }
         }
 

src/Common/HashTable/Hash.h

@@ -1,9 +1,10 @@
 #pragma once
 
+#include <city.h>
+#include <Core/Types.h>
 #include <common/types.h>
 #include <common/unaligned.h>
 #include <common/StringRef.h>
-#include <Common/UInt128.h>
 
 #include <type_traits>
 
@@ -167,35 +168,36 @@ inline size_t DefaultHash64(std::enable_if_t<(sizeof(T) <= sizeof(UInt64)), T> k
     return intHash64(u.out);
 }
 
-template <typename T>
-static constexpr bool UseDefaultHashForBigInts =
-    std::is_same_v<T, DB::Int128>  ||
-    std::is_same_v<T, DB::UInt128> ||
-    (is_big_int_v<T> && sizeof(T) == 32);
 
 template <typename T>
-inline size_t DefaultHash64(std::enable_if_t<(sizeof(T) > sizeof(UInt64) && UseDefaultHashForBigInts<T>), T> key)
+inline size_t DefaultHash64(std::enable_if_t<(sizeof(T) > sizeof(UInt64)), T> key)
 {
-    if constexpr (std::is_same_v<T, DB::Int128>)
+    if constexpr (is_big_int_v<T> && sizeof(T) == 16)
     {
-        return intHash64(static_cast<UInt64>(key) ^ static_cast<UInt64>(key >> 64));
+        /// TODO This is classical antipattern.
+        return intHash64(
+            static_cast<UInt64>(key) ^
+            static_cast<UInt64>(key >> 64));
     }
-    else if constexpr (std::is_same_v<T, DB::UInt128>)
+    else if constexpr (std::is_same_v<T, DB::UUID>)
     {
-        return intHash64(key.low ^ key.high);
+        return intHash64(
+            static_cast<UInt64>(key.toUnderType()) ^
+            static_cast<UInt64>(key.toUnderType() >> 64));
     }
     else if constexpr (is_big_int_v<T> && sizeof(T) == 32)
     {
-        return intHash64(static_cast<UInt64>(key) ^
+        return intHash64(
+            static_cast<UInt64>(key) ^
             static_cast<UInt64>(key >> 64) ^
             static_cast<UInt64>(key >> 128) ^
             static_cast<UInt64>(key >> 256));
     }
-
     assert(false);
     __builtin_unreachable();
 }
 
+
 template <typename T, typename Enable = void>
 struct DefaultHash;
 
@@ -213,7 +215,7 @@ struct DefaultHash<T, std::enable_if_t<DB::IsDecimalNumber<T>>>
 {
     size_t operator() (T key) const
     {
-        return DefaultHash64<typename T::NativeType>(key);
+        return DefaultHash64<typename T::NativeType>(key.value);
     }
 };
 
@@ -261,15 +263,98 @@ DEFINE_HASH(DB::Int128)
 DEFINE_HASH(DB::Int256)
 DEFINE_HASH(DB::Float32)
 DEFINE_HASH(DB::Float64)
+DEFINE_HASH(DB::UUID)
 
 #undef DEFINE_HASH
 
 
-template <>
-struct DefaultHash<DB::UInt128> : public DB::UInt128Hash {};
+struct UInt128Hash
+{
+    size_t operator()(UInt128 x) const
+    {
+        return CityHash_v1_0_2::Hash128to64({x.items[0], x.items[1]});
+    }
+};
+
+struct UUIDHash
+{
+    size_t operator()(DB::UUID x) const
+    {
+        return UInt128Hash()(x.toUnderType());
+    }
+};
+
+#ifdef __SSE4_2__
+
+struct UInt128HashCRC32
+{
+    size_t operator()(UInt128 x) const
+    {
+        UInt64 crc = -1ULL;
+        crc = _mm_crc32_u64(crc, x.items[0]);
+        crc = _mm_crc32_u64(crc, x.items[1]);
+        return crc;
+    }
+};
+
+#else
+
+/// On other platforms we do not use CRC32. NOTE This can be confusing.
+struct UInt128HashCRC32 : public UInt128Hash {};
+
+#endif
+
+struct UInt128TrivialHash
+{
+    size_t operator()(UInt128 x) const { return x.items[0]; }
+};
+
+struct UUIDTrivialHash
+{
+    size_t operator()(DB::UUID x) const { return x.toUnderType().items[0]; }
+};
+
+struct UInt256Hash
+{
+    size_t operator()(UInt256 x) const
+    {
+        /// NOTE suboptimal
+        return CityHash_v1_0_2::Hash128to64({
+            CityHash_v1_0_2::Hash128to64({x.items[0], x.items[1]}),
+            CityHash_v1_0_2::Hash128to64({x.items[2], x.items[3]})});
+    }
+};
+
+#ifdef __SSE4_2__
+
+struct UInt256HashCRC32
+{
+    size_t operator()(UInt256 x) const
+    {
+        UInt64 crc = -1ULL;
+        crc = _mm_crc32_u64(crc, x.items[0]);
+        crc = _mm_crc32_u64(crc, x.items[1]);
+        crc = _mm_crc32_u64(crc, x.items[2]);
+        crc = _mm_crc32_u64(crc, x.items[3]);
+        return crc;
+    }
+};
+
+#else
+
+/// We do not need to use CRC32 on other platforms. NOTE This can be confusing.
+struct UInt256HashCRC32 : public UInt256Hash {};
+
+#endif
 
 template <>
-struct DefaultHash<DB::DummyUInt256> : public DB::UInt256Hash {};
+struct DefaultHash<DB::UInt128> : public UInt128Hash {};
+
+template <>
+struct DefaultHash<DB::UInt256> : public UInt256Hash {};
+
+template <>
+struct DefaultHash<DB::UUID> : public UUIDHash {};
 
 
 /// It is reasonable to use for UInt8, UInt16 with sufficient hash table size.
@@ -322,23 +407,18 @@ struct IntHash32
 {
     size_t operator() (const T & key) const
     {
-        if constexpr (std::is_same_v<T, DB::Int128>)
+        if constexpr (is_big_int_v<T> && sizeof(T) == 16)
         {
-            return intHash32<salt>(static_cast<UInt64>(key) ^ static_cast<UInt64>(key >> 64));
-        }
-        else if constexpr (std::is_same_v<T, DB::UInt128>)
-        {
-            return intHash32<salt>(key.low ^ key.high);
+            return intHash32<salt>(key.items[0] ^ key.items[1]);
         }
         else if constexpr (is_big_int_v<T> && sizeof(T) == 32)
         {
-            return intHash32<salt>(static_cast<UInt64>(key) ^
-                static_cast<UInt64>(key >> 64) ^
-                static_cast<UInt64>(key >> 128) ^
-                static_cast<UInt64>(key >> 256));
+            return intHash32<salt>(key.items[0] ^ key.items[1] ^ key.items[2] ^ key.items[3]);
         }
         else if constexpr (sizeof(T) <= sizeof(UInt64))
+        {
             return intHash32<salt>(key);
+        }
 
         assert(false);
         __builtin_unreachable();

src/Common/HashTable/HashTable.h

@@ -70,10 +70,10 @@ namespace ZeroTraits
 {
 
 template <typename T>
-bool check(const T x) { return x == 0; }
+bool check(const T x) { return x == T{}; }
 
 template <typename T>
-void set(T & x) { x = 0; }
+void set(T & x) { x = {}; }
 
 }
 

src/Common/HashTable/StringHashMap.h

@@ -28,9 +28,8 @@ struct StringHashMapCell<StringKey16, TMapped> : public HashMapCell<StringKey16,
 
     // Zero means unoccupied cells in hash table. Use key with last word = 0 as
     // zero keys, because such keys are unrepresentable (no way to encode length).
-    static bool isZero(const StringKey16 & key, const HashTableNoState &)
-    { return key.high == 0; }
-    void setZero() { this->value.first.high = 0; }
+    static bool isZero(const StringKey16 & key, const HashTableNoState &) { return key.items[1] == 0; }
+    void setZero() { this->value.first.items[1] = 0; }
 
     // external
     const StringRef getKey() const { return toStringRef(this->value.first); }

src/Common/HashTable/StringHashSet.h

@@ -33,8 +33,8 @@ struct StringHashSetCell<StringKey16> : public HashTableCell<StringKey16, String
     // Zero means unoccupied cells in hash table. Use key with last word = 0 as
     // zero keys, because such keys are unrepresentable (no way to encode length).
     static bool isZero(const StringKey16 & key_, const HashTableNoState &)
-    { return key_.high == 0; }
-    void setZero() { this->key.high = 0; }
+    { return key_.items[1] == 0; }
+    void setZero() { this->key.items[1] = 0; }
 };
 
 template <>

src/Common/HashTable/StringHashTable.h

@@ -25,8 +25,8 @@ inline StringRef ALWAYS_INLINE toStringRef(const StringKey8 & n)
 }
 inline StringRef ALWAYS_INLINE toStringRef(const StringKey16 & n)
 {
-    assert(n.high != 0);
-    return {reinterpret_cast<const char *>(&n), 16ul - (__builtin_clzll(n.high) >> 3)};
+    assert(n.items[1] != 0);
+    return {reinterpret_cast<const char *>(&n), 16ul - (__builtin_clzll(n.items[1]) >> 3)};
 }
 inline StringRef ALWAYS_INLINE toStringRef(const StringKey24 & n)
 {
@@ -46,8 +46,8 @@ struct StringHashTableHash
     size_t ALWAYS_INLINE operator()(StringKey16 key) const
     {
         size_t res = -1ULL;
-        res = _mm_crc32_u64(res, key.low);
-        res = _mm_crc32_u64(res, key.high);
+        res = _mm_crc32_u64(res, key.items[0]);
+        res = _mm_crc32_u64(res, key.items[1]);
         return res;
     }
     size_t ALWAYS_INLINE operator()(StringKey24 key) const

src/Common/MemoryTracker.cpp

@@ -6,12 +6,14 @@
 #include <Common/formatReadable.h>
 #include <common/logger_useful.h>
 #include <Common/ProfileEvents.h>
+#include <Common/thread_local_rng.h>
 
 #include <atomic>
 #include <cmath>
 #include <random>
 #include <cstdlib>
 
+
 #ifdef MEMORY_TRACKER_DEBUG_CHECKS
 thread_local bool _memory_tracker_always_throw_logical_error_on_allocation = false;
 #endif

src/Common/OpenTelemetryTraceContext.h

@@ -6,12 +6,12 @@ namespace DB
 // The runtime info we need to create new OpenTelemetry spans.
 struct OpenTelemetryTraceContext
 {
-    __uint128_t trace_id = 0;
+    UUID trace_id{};
     UInt64 span_id = 0;
     // The incoming tracestate header and the trace flags, we just pass them
     // downstream. See https://www.w3.org/TR/trace-context/
     String tracestate;
-    uint8_t trace_flags = 0;
+    UInt8 trace_flags = 0;
 
     // Parse/compose OpenTelemetry traceparent header.
     bool parseTraceparentHeader(const std::string & traceparent, std::string & error);

src/Common/PODArray.h

@@ -260,22 +260,12 @@ public:
     template <typename ... TAllocatorParams>
     void push_back_raw(const void * ptr, TAllocatorParams &&... allocator_params)
     {
-        push_back_raw_many(1, ptr, std::forward<TAllocatorParams>(allocator_params)...);
-    }
-
-    template <typename ... TAllocatorParams>
-    void push_back_raw_many(size_t number_of_items, const void * ptr, TAllocatorParams &&... allocator_params)
-    {
-        size_t required_capacity = size() + number_of_items;
+        size_t required_capacity = size() + ELEMENT_SIZE;
         if (unlikely(required_capacity > capacity()))
             reserve(required_capacity, std::forward<TAllocatorParams>(allocator_params)...);
 
-        size_t items_byte_size = byte_size(number_of_items);
-        if (items_byte_size)
-        {
-            memcpy(c_end, ptr, items_byte_size);
-            c_end += items_byte_size;
-        }
+        memcpy(c_end, ptr, ELEMENT_SIZE);
+        c_end += ELEMENT_SIZE;
     }
 
     void protect()

src/Common/SipHash.h

@@ -155,7 +155,6 @@ public:
         unalignedStore<UInt64>(out + 8, v2 ^ v3);
     }
 
-    /// template for avoiding 'unsigned long long' vs 'unsigned long' problem on old poco in macos
     template <typename T>
     ALWAYS_INLINE void get128(T & lo, T & hi)
     {
@@ -165,6 +164,13 @@ public:
         hi = v2 ^ v3;
     }
 
+    template <typename T>
+    ALWAYS_INLINE void get128(T & dst)
+    {
+        static_assert(sizeof(T) == 16);
+        get128(reinterpret_cast<char *>(&dst));
+    }
+
     UInt64 get64()
     {
         finalize();

src/Common/SpaceSaving.h

@@ -6,7 +6,6 @@
 #include <boost/range/adaptor/reversed.hpp>
 
 #include <Common/ArenaWithFreeLists.h>
-#include <Common/UInt128.h>
 #include <Common/HashTable/Hash.h>
 #include <Common/HashTable/HashMap.h>
 

src/Common/TraceCollector.cpp

@@ -24,7 +24,10 @@ namespace
     /// Thus upper bound on query_id length should be introduced to avoid buffer overflow in signal handler.
     ///
     /// And it cannot be large, since otherwise it will not fit into PIPE_BUF.
-    constexpr size_t QUERY_ID_MAX_LEN = sizeof("00000000-0000-0000-0000-000000000000") - 1; // 36
+    /// The performance test query ids can be surprisingly long like
+    /// `aggregating_merge_tree_simple_aggregate_function_string.query100.profile100`,
+    /// so make some allowance for them as well.
+    constexpr size_t QUERY_ID_MAX_LEN = 128;
 }
 
 LazyPipeFDs pipe;

src/Common/UInt128.h

@@ -1,292 +0,0 @@
-#pragma once
-
-#include <tuple>
-#include <iomanip>
-#include <city.h>
-
-#include <Core/Types.h>
-#include <Common/hex.h>
-
-#ifdef __SSE4_2__
-#include <nmmintrin.h>
-#endif
-
-
-namespace DB
-{
-
-/// For aggregation by SipHash, UUID type or concatenation of several fields.
-struct UInt128
-{
-/// Suppress gcc7 warnings: 'prev_key.DB::UInt128::low' may be used uninitialized in this function
-#if !defined(__clang__)
-#pragma GCC diagnostic push
-#pragma GCC diagnostic ignored "-Wmaybe-uninitialized"
-#endif
-
-    /// This naming assumes little endian.
-    UInt64 low;
-    UInt64 high;
-
-    /// TODO: Make this constexpr. Currently it is used in unions
-    /// and union cannot contain member with non trivial constructor
-    /// constructor must be non user provided but compiler cannot constexpr constructor
-    /// if members low and high are not initialized, if we default member initialize them
-    /// constructor becomes non trivial.
-    UInt128() = default;
-    explicit constexpr UInt128(const UInt64 low_, const UInt64 high_) : low(low_), high(high_) { }
-
-    /// We need Int128 to UInt128 conversion or AccurateComparison will call greaterOp<Int128, UInt64> instead of greaterOp<Int128, UInt128>
-    explicit constexpr UInt128(const Int128 rhs) : low(rhs), high(rhs >> 64) {}
-    explicit constexpr UInt128(const Int64 rhs) : low(rhs), high() {}
-    explicit constexpr UInt128(const Int32 rhs) : low(rhs), high() {}
-    explicit constexpr UInt128(const Int16 rhs) : low(rhs), high() {}
-    explicit constexpr UInt128(const Int8 rhs) : low(rhs), high() {}
-    explicit constexpr UInt128(const UInt8 rhs) : low(rhs), high() {}
-    explicit constexpr UInt128(const UInt16 rhs) : low(rhs), high() {}
-    explicit constexpr UInt128(const UInt32 rhs) : low(rhs), high() {}
-    explicit constexpr UInt128(const UInt64 rhs) : low(rhs), high() {}
-    explicit constexpr UInt128(const Float32 rhs) : low(rhs), high() {}
-    explicit constexpr UInt128(const Float64 rhs) : low(rhs), high() {}
-
-    constexpr auto tuple() const { return std::tie(high, low); }
-
-    String toHexString() const
-    {
-        String res(2 * sizeof(UInt128), 0);
-        writeHexUIntLowercase(*this, res.data());
-        return res;
-    }
-
-    constexpr bool operator== (const UInt128 rhs) const { return tuple() == rhs.tuple(); }
-    constexpr bool operator!= (const UInt128 rhs) const { return tuple() != rhs.tuple(); }
-    constexpr bool operator<  (const UInt128 rhs) const { return tuple() < rhs.tuple(); }
-    constexpr bool operator<= (const UInt128 rhs) const { return tuple() <= rhs.tuple(); }
-    constexpr bool operator>  (const UInt128 rhs) const { return tuple() > rhs.tuple(); }
-    constexpr bool operator>= (const UInt128 rhs) const { return tuple() >= rhs.tuple(); }
-
-    constexpr bool operator == (const Int128 rhs) const { return *this == UInt128(rhs, rhs >> 64) && rhs >= 0; }
-    constexpr bool operator != (const Int128 rhs) const { return *this != UInt128(rhs, rhs >> 64) || rhs < 0; }
-    constexpr bool operator >= (const Int128 rhs) const { return *this >= UInt128(rhs, rhs >> 64) || rhs < 0; }
-    constexpr bool operator >  (const Int128 rhs) const { return *this >  UInt128(rhs, rhs >> 64) || rhs < 0; }
-    constexpr bool operator <= (const Int128 rhs) const { return *this <= UInt128(rhs, rhs >> 64) && rhs >= 0; }
-    constexpr bool operator <  (const Int128 rhs) const { return *this <  UInt128(rhs, rhs >> 64) && rhs >= 0; }
-
-    constexpr bool operator >  (const Int256 rhs) const { return (rhs < 0) || ((Int256(high) << 64) + low) > rhs; }
-    constexpr bool operator >  (const UInt256 rhs) const { return ((UInt256(high) << 64) + low) > rhs; }
-    constexpr bool operator <  (const Int256 rhs) const { return (rhs >= 0) && ((Int256(high) << 64) + low) < rhs; }
-    constexpr bool operator <  (const UInt256 rhs) const { return ((UInt256(high) << 64) + low) < rhs; }
-
-    template <typename T> constexpr bool operator== (const T rhs) const { return *this == UInt128(rhs); }
-    template <typename T> constexpr bool operator!= (const T rhs) const { return *this != UInt128(rhs); }
-    template <typename T> constexpr bool operator>= (const T rhs) const { return *this >= UInt128(rhs); }
-    template <typename T> constexpr bool operator>  (const T rhs) const { return *this >  UInt128(rhs); }
-    template <typename T> constexpr bool operator<= (const T rhs) const { return *this <= UInt128(rhs); }
-    template <typename T> constexpr bool operator<  (const T rhs) const { return *this <  UInt128(rhs); }
-
-    template <typename T> explicit operator T() const
-    {
-        if constexpr (std::is_class_v<T>)
-            return T();
-        else
-            return static_cast<T>(low);
-    }
-
-#if !defined(__clang__)
-#pragma GCC diagnostic pop
-#endif
-
-    constexpr UInt128 & operator= (const UInt64 rhs) { low = rhs; high = 0; return *this; }
-};
-
-template <typename T> constexpr bool operator == (T a, const UInt128 b) { return b.operator==(a); }
-template <typename T> constexpr bool operator != (T a, const UInt128 b) { return b.operator!=(a); }
-template <typename T> constexpr bool operator >= (T a, const UInt128 b) { return b <= a; }
-template <typename T> constexpr bool operator >  (T a, const UInt128 b) { return b < a; }
-template <typename T> constexpr bool operator <= (T a, const UInt128 b) { return b >= a; }
-template <typename T> constexpr bool operator <  (T a, const UInt128 b) { return b > a; }
-
-template <> inline constexpr bool IsNumber<UInt128> = true;
-template <> struct TypeName<UInt128> { static constexpr const char * get() { return "UInt128"; } };
-template <> struct TypeId<UInt128> { static constexpr const TypeIndex value = TypeIndex::UInt128; };
-
-struct UInt128Hash
-{
-    size_t operator()(UInt128 x) const
-    {
-        return CityHash_v1_0_2::Hash128to64({x.low, x.high});
-    }
-};
-
-#ifdef __SSE4_2__
-
-struct UInt128HashCRC32
-{
-    size_t operator()(UInt128 x) const
-    {
-        UInt64 crc = -1ULL;
-        crc = _mm_crc32_u64(crc, x.low);
-        crc = _mm_crc32_u64(crc, x.high);
-        return crc;
-    }
-};
-
-#else
-
-/// On other platforms we do not use CRC32. NOTE This can be confusing.
-struct UInt128HashCRC32 : public UInt128Hash {};
-
-#endif
-
-struct UInt128TrivialHash
-{
-    size_t operator()(UInt128 x) const { return x.low; }
-};
-
-
-/** Used for aggregation, for putting a large number of constant-length keys in a hash table.
-  */
-struct DummyUInt256
-{
-
-/// Suppress gcc7 warnings: 'prev_key.DB::UInt256::a' may be used uninitialized in this function
-#if !defined(__clang__)
-#pragma GCC diagnostic push
-#pragma GCC diagnostic ignored "-Wmaybe-uninitialized"
-#endif
-
-    UInt64 a;
-    UInt64 b;
-    UInt64 c;
-    UInt64 d;
-
-    bool operator== (const DummyUInt256 rhs) const
-    {
-        return a == rhs.a && b == rhs.b && c == rhs.c && d == rhs.d;
-
-    /* So it's no better.
-        return 0xFFFF == _mm_movemask_epi8(_mm_and_si128(
-            _mm_cmpeq_epi8(
-                _mm_loadu_si128(reinterpret_cast<const __m128i *>(&a)),
-                _mm_loadu_si128(reinterpret_cast<const __m128i *>(&rhs.a))),
-            _mm_cmpeq_epi8(
-                _mm_loadu_si128(reinterpret_cast<const __m128i *>(&c)),
-                _mm_loadu_si128(reinterpret_cast<const __m128i *>(&rhs.c)))));*/
-    }
-
-    bool operator!= (const DummyUInt256 rhs) const { return !operator==(rhs); }
-
-    bool operator== (const UInt64 rhs) const { return a == rhs && b == 0 && c == 0 && d == 0; }
-    bool operator!= (const UInt64 rhs) const { return !operator==(rhs); }
-
-#if !defined(__clang__)
-#pragma GCC diagnostic pop
-#endif
-
-    DummyUInt256 & operator = (const UInt64 rhs) { a = rhs; b = 0; c = 0; d = 0; return *this; }
-};
-
-struct UInt256Hash
-{
-    size_t operator()(DummyUInt256 x) const
-    {
-        /// NOTE suboptimal
-        return CityHash_v1_0_2::Hash128to64({CityHash_v1_0_2::Hash128to64({x.a, x.b}), CityHash_v1_0_2::Hash128to64({x.c, x.d})});
-    }
-};
-
-#ifdef __SSE4_2__
-
-struct UInt256HashCRC32
-{
-    size_t operator()(DummyUInt256 x) const
-    {
-        UInt64 crc = -1ULL;
-        crc = _mm_crc32_u64(crc, x.a);
-        crc = _mm_crc32_u64(crc, x.b);
-        crc = _mm_crc32_u64(crc, x.c);
-        crc = _mm_crc32_u64(crc, x.d);
-        return crc;
-    }
-};
-
-#else
-
-/// We do not need to use CRC32 on other platforms. NOTE This can be confusing.
-struct UInt256HashCRC32 : public UInt256Hash {};
-
-#endif
-
-}
-
-template <> struct is_signed<DB::UInt128>
-{
-    static constexpr bool value = false;
-};
-
-template <> struct is_unsigned<DB::UInt128>
-{
-    static constexpr bool value = true;
-};
-
-template <> struct is_integer<DB::UInt128>
-{
-    static constexpr bool value = true;
-};
-
-// Operator +, -, /, *, % aren't implemented so it's not an arithmetic type
-template <> struct is_arithmetic<DB::UInt128>
-{
-    static constexpr bool value = false;
-};
-
-/// Overload hash for type casting
-namespace std
-{
-template <> struct hash<DB::UInt128>
-{
-    size_t operator()(const DB::UInt128 & u) const
-    {
-        return CityHash_v1_0_2::Hash128to64({u.low, u.high});
-    }
-};
-
-template<>
-class numeric_limits<DB::UInt128>
-{
-public:
-    static constexpr bool is_specialized = true;
-    static constexpr bool is_signed = ::is_signed<DB::UInt128>::value;
-    static constexpr bool is_integer = ::is_integer<DB::UInt128>::value;
-    static constexpr bool is_exact = true;
-    static constexpr bool has_infinity = false;
-    static constexpr bool has_quiet_NaN = false;
-    static constexpr bool has_signaling_NaN = false;
-    static constexpr std::float_denorm_style has_denorm = std::denorm_absent;
-    static constexpr bool has_denorm_loss = false;
-    static constexpr std::float_round_style round_style = std::round_toward_zero;
-    static constexpr bool is_iec559 = false;
-    static constexpr bool is_bounded = true;
-    static constexpr bool is_modulo = true;
-    static constexpr int digits = std::numeric_limits<UInt64>::digits * 2;
-    static constexpr int digits10 = digits * 0.30103 /*std::log10(2)*/;
-    static constexpr int max_digits10 = 0;
-    static constexpr int radix = 2;
-    static constexpr int min_exponent = 0;
-    static constexpr int min_exponent10 = 0;
-    static constexpr int max_exponent = 0;
-    static constexpr int max_exponent10 = 0;
-    static constexpr bool traps = true;
-    static constexpr bool tinyness_before = false;
-
-    static constexpr DB::UInt128 min() noexcept { return DB::UInt128(0, 0); }
-
-    static constexpr DB::UInt128 max() noexcept
-    {
-        return DB::UInt128(std::numeric_limits<UInt64>::max(), std::numeric_limits<UInt64>::max());
-    }
-
-    static constexpr DB::UInt128 lowest() noexcept { return min(); }
-};
-
-}

src/Common/config.h.in

@@ -15,3 +15,4 @@
 #cmakedefine01 USE_GRPC
 #cmakedefine01 USE_STATS
 #cmakedefine01 CLICKHOUSE_SPLIT_BINARY
+#cmakedefine01 USE_DATASKETCHES

src/Common/examples/arena_with_free_lists.cpp

@@ -135,24 +135,24 @@ struct Dictionary
     template <typename Value> using ContainerType = Value[];
     template <typename Value> using ContainerPtrType = std::unique_ptr<ContainerType<Value>>;
 
-    enum class AttributeUnderlyingType
+    enum class AttributeUnderlyingTypeTest
     {
-        utUInt8,
-        utUInt16,
-        utUInt32,
-        utUInt64,
-        utInt8,
-        utInt16,
-        utInt32,
-        utInt64,
-        utFloat32,
-        utFloat64,
-        utString
+        UInt8,
+        UInt16,
+        UInt32,
+        UInt64,
+        Int8,
+        Int16,
+        Int32,
+        Int64,
+        Float32,
+        Float64,
+        String
     };
 
     struct Attribute final
     {
-        AttributeUnderlyingType type;
+        AttributeUnderlyingTypeTest type;
         std::variant<
             UInt8, UInt16, UInt32, UInt64,
             Int8, Int16, Int32, Int64,
@@ -172,17 +172,17 @@ struct Dictionary
     {
         switch (attribute.type)
         {
-            case AttributeUnderlyingType::utUInt8: std::get<ContainerPtrType<UInt8>>(attribute.arrays)[idx] = value.get<UInt64>(); break;
-            case AttributeUnderlyingType::utUInt16: std::get<ContainerPtrType<UInt16>>(attribute.arrays)[idx] = value.get<UInt64>(); break;
-            case AttributeUnderlyingType::utUInt32: std::get<ContainerPtrType<UInt32>>(attribute.arrays)[idx] = value.get<UInt64>(); break;
-            case AttributeUnderlyingType::utUInt64: std::get<ContainerPtrType<UInt64>>(attribute.arrays)[idx] = value.get<UInt64>(); break;
-            case AttributeUnderlyingType::utInt8: std::get<ContainerPtrType<Int8>>(attribute.arrays)[idx] = value.get<Int64>(); break;
-            case AttributeUnderlyingType::utInt16: std::get<ContainerPtrType<Int16>>(attribute.arrays)[idx] = value.get<Int64>(); break;
-            case AttributeUnderlyingType::utInt32: std::get<ContainerPtrType<Int32>>(attribute.arrays)[idx] = value.get<Int64>(); break;
-            case AttributeUnderlyingType::utInt64: std::get<ContainerPtrType<Int64>>(attribute.arrays)[idx] = value.get<Int64>(); break;
-            case AttributeUnderlyingType::utFloat32: std::get<ContainerPtrType<Float32>>(attribute.arrays)[idx] = value.get<Float64>(); break;
-            case AttributeUnderlyingType::utFloat64: std::get<ContainerPtrType<Float64>>(attribute.arrays)[idx] = value.get<Float64>(); break;
-            case AttributeUnderlyingType::utString:
+            case AttributeUnderlyingTypeTest::UInt8: std::get<ContainerPtrType<UInt8>>(attribute.arrays)[idx] = value.get<UInt64>(); break;
+            case AttributeUnderlyingTypeTest::UInt16: std::get<ContainerPtrType<UInt16>>(attribute.arrays)[idx] = value.get<UInt64>(); break;
+            case AttributeUnderlyingTypeTest::UInt32: std::get<ContainerPtrType<UInt32>>(attribute.arrays)[idx] = value.get<UInt64>(); break;
+            case AttributeUnderlyingTypeTest::UInt64: std::get<ContainerPtrType<UInt64>>(attribute.arrays)[idx] = value.get<UInt64>(); break;
+            case AttributeUnderlyingTypeTest::Int8: std::get<ContainerPtrType<Int8>>(attribute.arrays)[idx] = value.get<Int64>(); break;
+            case AttributeUnderlyingTypeTest::Int16: std::get<ContainerPtrType<Int16>>(attribute.arrays)[idx] = value.get<Int64>(); break;
+            case AttributeUnderlyingTypeTest::Int32: std::get<ContainerPtrType<Int32>>(attribute.arrays)[idx] = value.get<Int64>(); break;
+            case AttributeUnderlyingTypeTest::Int64: std::get<ContainerPtrType<Int64>>(attribute.arrays)[idx] = value.get<Int64>(); break;
+            case AttributeUnderlyingTypeTest::Float32: std::get<ContainerPtrType<Float32>>(attribute.arrays)[idx] = value.get<Float64>(); break;
+            case AttributeUnderlyingTypeTest::Float64: std::get<ContainerPtrType<Float64>>(attribute.arrays)[idx] = value.get<Float64>(); break;
+            case AttributeUnderlyingTypeTest::String:
             {
                 const auto & string = value.get<String>();
                 auto & string_ref = std::get<ContainerPtrType<StringRef>>(attribute.arrays)[idx];
@@ -308,7 +308,7 @@ int main(int argc, char ** argv)
     constexpr size_t cache_size = 1024;
 
     Dictionary::Attribute attr;
-    attr.type = Dictionary::AttributeUnderlyingType::utString;
+    attr.type = Dictionary::AttributeUnderlyingTypeTest::String;
     std::get<Dictionary::ContainerPtrType<StringRef>>(attr.arrays).reset(new StringRef[cache_size]{});  // NOLINT
 
     while (true)

src/Common/getHashOfLoadedBinary.cpp

@@ -38,7 +38,7 @@ std::string getHashOfLoadedBinaryHex()
 {
     SipHash hash = getHashOfLoadedBinary();
     std::array<UInt64, 2> checksum;
-    hash.get128(checksum[0], checksum[1]);
+    hash.get128(checksum);
     return getHexUIntUppercase(checksum);
 }
 

src/Common/intExp.h

@@ -138,9 +138,9 @@ constexpr inline Int128 exp10_i128(int x)
 }
 
 
-inline wInt256 exp10_i256(int x)
+inline Int256 exp10_i256(int x)
 {
-    using Int256 = wInt256;
+    using Int256 = Int256;
     static constexpr Int256 i10e18{1000000000000000000ll};
     static const Int256 values[] = {
         static_cast<Int256>(1ll),

src/Common/tests/gtest_hash_table.cpp

@@ -5,11 +5,17 @@
 
 #include <Common/HashTable/HashMap.h>
 #include <Common/HashTable/HashSet.h>
+#include <Common/HashTable/Hash.h>
 
 #include <IO/ReadBufferFromString.h>
+#include <IO/WriteHelpers.h>
 
 #include <gtest/gtest.h>
 
+
+using namespace DB;
+
+
 /// To test dump functionality without using other hashes that can change
 template <typename T>
 struct DummyHash
@@ -18,12 +24,12 @@ struct DummyHash
 };
 
 template<typename HashTable>
-std::set<typename HashTable::value_type> convertToSet(const HashTable& table)
+std::set<std::string> convertToSet(const HashTable & table)
 {
-    std::set<typename HashTable::value_type> result;
+    std::set<std::string> result;
 
     for (auto v: table)
-        result.emplace(v.getValue());
+        result.emplace(toString(v.getValue()));
 
     return result;
 }
@@ -91,8 +97,8 @@ TEST(HashTable, Iteration)
     cont.insert(2);
     cont.insert(3);
 
-    std::set<int> expected = {1, 2, 3};
-    std::set<int> actual = convertToSet(cont);
+    std::set<std::string> expected = {"1", "2", "3"};
+    std::set<std::string> actual = convertToSet(cont);
 
     ASSERT_EQ(actual, expected);
 }
@@ -251,14 +257,14 @@ TEST(HashTable, SerializationDeserialization)
         cont.insert(2);
         cont.insert(3);
 
-        DB::WriteBufferFromOwnString wb;
+        WriteBufferFromOwnString wb;
         cont.writeText(wb);
 
         std::string expected = "3,1,2,3";
 
         ASSERT_EQ(wb.str(), expected);
 
-        DB::ReadBufferFromString rb(expected);
+        ReadBufferFromString rb(expected);
 
         Cont deserialized;
         deserialized.readText(rb);
@@ -273,10 +279,10 @@ TEST(HashTable, SerializationDeserialization)
         cont.insert(2);
         cont.insert(3);
 
-        DB::WriteBufferFromOwnString wb;
+        WriteBufferFromOwnString wb;
         cont.write(wb);
 
-        DB::ReadBufferFromString rb(wb.str());
+        ReadBufferFromString rb(wb.str());
 
         Cont deserialized;
         deserialized.read(rb);
@@ -286,23 +292,23 @@ TEST(HashTable, SerializationDeserialization)
         using Cont = HashSet<int, DummyHash<int>, HashTableGrower<1>>;
         Cont cont;
 
-        DB::WriteBufferFromOwnString wb;
+        WriteBufferFromOwnString wb;
         cont.writeText(wb);
 
         std::string expected = "0";
         ASSERT_EQ(wb.str(), expected);
 
-        DB::ReadBufferFromString rb(expected);
+        ReadBufferFromString rb(expected);
 
         Cont deserialized;
         deserialized.readText(rb);
         ASSERT_EQ(convertToSet(cont), convertToSet(deserialized));
     }
     {
-        using Cont = HashSet<DB::UInt128, DB::UInt128TrivialHash>;
+        using Cont = HashSet<UInt128, UInt128TrivialHash>;
         Cont cont;
 
-        DB::WriteBufferFromOwnString wb;
+        WriteBufferFromOwnString wb;
         cont.write(wb);
 
         std::string expected;
@@ -310,7 +316,7 @@ TEST(HashTable, SerializationDeserialization)
 
         ASSERT_EQ(wb.str(), expected);
 
-        DB::ReadBufferFromString rb(expected);
+        ReadBufferFromString rb(expected);
 
         Cont deserialized;
         deserialized.read(rb);
@@ -359,8 +365,8 @@ TEST(HashTable, Resize)
         cont.insert(3);
         cont.insert(1);
 
-        std::set<int> expected = {1, 3};
-        std::set<int> actual = convertToSet(cont);
+        std::set<std::string> expected = {"1", "3"};
+        std::set<std::string> actual = convertToSet(cont);
 
         ASSERT_EQ(actual, expected);
     }

src/Common/tests/gtest_pod_array.cpp

@@ -459,18 +459,6 @@ TEST(Common, PODArrayInsertFromItself)
     }
 }
 
-TEST(Common, PODPushBackRawMany)
-{
-    PODArray<char> chars;
-    chars.push_back_raw_many(5, "first");
-    EXPECT_EQ(std::string("first"), std::string(chars.data(), chars.size()));
-    EXPECT_EQ(5, chars.size());
-    EXPECT_LE(chars.capacity() - chars.size(), 10);
-    chars.push_back_raw_many(10, "0123456789");
-    EXPECT_EQ(15, chars.size());
-    EXPECT_EQ(std::string("first0123456789"), std::string(chars.data(), chars.size()));
-}
-
 TEST(Common, PODNoOverallocation)
 {
     /// Check that PaddedPODArray allocates for smaller number of elements than the power of two due to padding.

src/Common/tests/gtest_wide_integer.cpp

@@ -0,0 +1,281 @@
+#include <gtest/gtest.h>
+
+#include <cstdint>
+#include <limits>
+#include <type_traits>
+#include <initializer_list>
+
+#include <Core/Types.h>
+#include <IO/WriteHelpers.h>
+#include <IO/ReadHelpers.h>
+#include <common/demangle.h>
+
+
+static_assert(is_signed_v<Int128>);
+static_assert(!is_unsigned_v<Int128>);
+static_assert(is_integer_v<Int128>);
+static_assert(sizeof(Int128) == 16);
+
+static_assert(is_signed_v<Int256>);
+static_assert(!is_unsigned_v<Int256>);
+static_assert(is_integer_v<Int256>);
+static_assert(sizeof(Int256) == 32);
+
+static_assert(!is_signed_v<UInt128>);
+static_assert(is_unsigned_v<UInt128>);
+static_assert(is_integer_v<UInt128>);
+static_assert(sizeof(UInt128) == 16);
+
+static_assert(!is_signed_v<UInt256>);
+static_assert(is_unsigned_v<UInt256>);
+static_assert(is_integer_v<UInt256>);
+static_assert(sizeof(UInt256) == 32);
+
+
+using namespace DB;
+
+
+GTEST_TEST(WideInteger, Conversions)
+{
+    ASSERT_EQ(toString(UInt128(12345678901234567890ULL)), "12345678901234567890");
+    ASSERT_EQ(toString(UInt256(12345678901234567890ULL)), "12345678901234567890");
+
+    Int128 minus_one = -1;
+    ASSERT_EQ(minus_one.items[0], -1);
+    ASSERT_EQ(minus_one.items[1], -1);
+
+    ASSERT_EQ(0, memcmp(&minus_one, "\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF", sizeof(minus_one)));
+
+    ASSERT_EQ(minus_one, -1);
+    ASSERT_EQ(minus_one, -1LL);
+    ASSERT_EQ(minus_one, Int8(-1));
+    ASSERT_EQ(minus_one, Int16(-1));
+    ASSERT_EQ(minus_one, Int32(-1));
+    ASSERT_EQ(minus_one, Int64(-1));
+
+    ASSERT_LT(minus_one, 0);
+
+    Int128 zero = 0;
+    zero += -1;
+    ASSERT_EQ(zero, -1);
+    ASSERT_EQ(zero, minus_one);
+
+    zero += minus_one;
+    ASSERT_EQ(0, memcmp(&zero, "\xFE\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF", sizeof(zero)));
+
+    zero += 2;
+    ASSERT_EQ(zero, 0);
+
+    ASSERT_EQ(toString(Int128(-1)), "-1");
+    ASSERT_EQ(toString(Int256(-1)), "-1");
+
+    ASSERT_EQ(toString(Int128(-1LL)), "-1");
+    ASSERT_EQ(toString(Int256(-1LL)), "-1");
+
+    ASSERT_EQ(toString(Int128(-1234567890123456789LL)), "-1234567890123456789");
+    ASSERT_EQ(toString(Int256(-1234567890123456789LL)), "-1234567890123456789");
+
+    ASSERT_EQ(UInt64(UInt128(12345678901234567890ULL)), 12345678901234567890ULL);
+    ASSERT_EQ(UInt64(UInt256(12345678901234567890ULL)), 12345678901234567890ULL);
+
+    ASSERT_EQ(__uint128_t(UInt128(12345678901234567890ULL)), 12345678901234567890ULL);
+    ASSERT_EQ(__uint128_t(UInt256(12345678901234567890ULL)), 12345678901234567890ULL);
+
+    ASSERT_EQ(__int128_t(Int128(-1234567890123456789LL)), -1234567890123456789LL);
+    ASSERT_EQ(__int128_t(Int256(-1234567890123456789LL)), -1234567890123456789LL);
+
+    ASSERT_EQ(toString(Int128(-1)), "-1");
+    ASSERT_EQ(toString(Int256(-1)), "-1");
+
+    ASSERT_EQ(toString(UInt128(123.456)), "123");
+    ASSERT_EQ(toString(UInt256(123.456)), "123");
+    ASSERT_EQ(toString(Int128(-123.456)), "-123");
+    ASSERT_EQ(toString(Int256(-123.456)), "-123");
+
+    ASSERT_EQ(toString(UInt128(123.456f)), "123");
+    ASSERT_EQ(toString(UInt256(123.456f)), "123");
+    ASSERT_EQ(toString(Int128(-123.456f)), "-123");
+    ASSERT_EQ(toString(Int256(-123.456f)), "-123");
+
+    ASSERT_EQ(toString(UInt128(1) * 1000000000 * 1000000000 * 1000000000 * 1000000000), "1000000000000000000000000000000000000");
+    ASSERT_EQ(Float64(UInt128(1) * 1000000000 * 1000000000 * 1000000000 * 1000000000), 1e36);
+
+    ASSERT_EQ(toString(UInt256(1) * 1000000000 * 1000000000 * 1000000000 * 1000000000 * 1000000000 * 1000000000 * 1000000000 * 1000000000),
+        "1000000000000000000000000000000000000000000000000000000000000000000000000");
+    ASSERT_EQ(Float64(UInt256(1) * 1000000000 * 1000000000 * 1000000000 * 1000000000 * 1000000000 * 1000000000 * 1000000000 * 1000000000), 1e72);
+
+    EXPECT_EQ(toString(parse<Int128>("148873535527910577765226390751398592640")), "148873535527910577765226390751398592640");
+    EXPECT_EQ(toString(parse<UInt128>("148873535527910577765226390751398592640")), "148873535527910577765226390751398592640");
+}
+
+
+template <typename T>
+static T divide(T & numerator, T && denominator)
+{
+    if (!denominator)
+        throwError("Division by zero");
+
+    T & n = numerator;
+    T & d = denominator;
+    T x = 1;
+    T quotient = 0;
+
+    /// Multiply d to the power of two until it will be greater than n.
+    /// The factor will be collected in x.
+    while (d <= n && ((d >> (sizeof(T) * 8 - 1)) & 1) == 0)
+    {
+        x <<= 1;
+        d <<= 1;
+    }
+
+    std::cerr << toString(x) << ", " << toString(d) << "\n";
+
+    while (x)
+    {
+        if (d <= n)
+        {
+            n -= d;
+            quotient |= x;
+        }
+
+        x >>= 1;
+        d >>= 1;
+    }
+
+    return quotient;
+}
+
+
+GTEST_TEST(WideInteger, Arithmetic)
+{
+    Int128 minus_one = -1;
+    Int128 zero = 0;
+
+    zero += -1;
+    ASSERT_EQ(zero, -1);
+    ASSERT_EQ(zero, minus_one);
+
+    zero += minus_one;
+    ASSERT_EQ(0, memcmp(&zero, "\xFE\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF", sizeof(zero)));
+
+    zero += 2;
+    ASSERT_EQ(zero, 0);
+
+    ASSERT_EQ(UInt256(12345678901234567890ULL) * 12345678901234567890ULL / 12345678901234567890ULL, 12345678901234567890ULL);
+    ASSERT_EQ(UInt256(12345678901234567890ULL) * UInt256(12345678901234567890ULL) / 12345678901234567890ULL, 12345678901234567890ULL);
+    ASSERT_EQ(UInt256(12345678901234567890ULL) * 12345678901234567890ULL / UInt256(12345678901234567890ULL), 12345678901234567890ULL);
+    ASSERT_EQ(UInt256(12345678901234567890ULL) * 12345678901234567890ULL / 12345678901234567890ULL, UInt256(12345678901234567890ULL));
+    ASSERT_EQ(UInt128(12345678901234567890ULL) * 12345678901234567890ULL / UInt128(12345678901234567890ULL), 12345678901234567890ULL);
+    ASSERT_EQ(UInt256(12345678901234567890ULL) * UInt128(12345678901234567890ULL) / 12345678901234567890ULL, 12345678901234567890ULL);
+
+    ASSERT_EQ(Int128(0) + Int32(-1), Int128(-1));
+
+    Int128 x(parse<Int128>("148873535527910577765226390751398592640"));
+    Int128 dividend = x / 10;
+    ASSERT_EQ(toString(dividend), "14887353552791057776522639075139859264");
+}
+
+
+GTEST_TEST(WideInteger, DecimalArithmetic)
+{
+    Decimal128 zero{};
+    Decimal32 addend = -1000;
+
+    zero += Decimal128(addend);
+    ASSERT_EQ(zero.value, -1000);
+
+    zero += addend;
+    ASSERT_EQ(zero.value, -2000);
+}
+
+
+GTEST_TEST(WideInteger, FromDouble)
+{
+    /// Check that we are being able to convert double to big integer without the help of floating point instructions.
+    /// (a prototype of a function that we may need)
+
+    double f = -123.456;
+    UInt64 u;
+    memcpy(&u, &f, sizeof(f));
+
+    bool is_negative = u >> 63;
+    uint16_t exponent = (u >> 52) & (((1ull << 12) - 1) >> 1);
+    int16_t normalized_exponent = exponent - 1023;
+    UInt64 mantissa = u & ((1ull << 52) - 1);
+
+    // std::cerr << is_negative << ", " << normalized_exponent << ", " << mantissa << "\n";
+
+    /// x = sign * (2 ^ normalized_exponent + mantissa * 2 ^ (normalized_exponent - mantissa_bits))
+
+    Int128 res = 0;
+
+    if (normalized_exponent >= 128)
+    {
+    }
+    else
+    {
+        res = mantissa;
+        if (normalized_exponent > 52)
+            res <<= (normalized_exponent - 52);
+        else
+            res >>= (52 - normalized_exponent);
+
+        if (normalized_exponent > 0)
+            res += Int128(1) << normalized_exponent;
+    }
+
+    if (is_negative)
+        res = -res;
+
+    ASSERT_EQ(toString(res), "-123");
+}
+
+
+GTEST_TEST(WideInteger, Shift)
+{
+    Int128 x = 1;
+
+    auto y = x << 64;
+    ASSERT_EQ(0, memcmp(&y, "\x00\x00\x00\x00\x00\x00\x00\x00\x01\x00\x00\x00\x00\x00\x00\x00", sizeof(Int128)));
+
+    auto z = y << 11;
+    ASSERT_EQ(toString(z), "37778931862957161709568");
+
+    auto a = x << 11;
+    ASSERT_EQ(a, 2048);
+
+    z >>= 64;
+    ASSERT_EQ(z, a);
+
+    x = -1;
+    y = x << 16;
+
+    ASSERT_EQ(0, memcmp(&y, "\x00\x00\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF", sizeof(Int128)));
+
+    y >>= 16;
+    ASSERT_EQ(0, memcmp(&y, "\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF", sizeof(Int128)));
+
+    y <<= 64;
+    ASSERT_EQ(0, memcmp(&y, "\x00\x00\x00\x00\x00\x00\x00\x00\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF", sizeof(Int128)));
+
+    y >>= 32;
+    ASSERT_EQ(0, memcmp(&y, "\x00\x00\x00\x00\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF\xFF", sizeof(Int128)));
+
+    y <<= 64;
+    ASSERT_EQ(0, memcmp(&y, "\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\xFF\xFF\xFF\xFF", sizeof(Int128)));
+}
+
+
+GTEST_TEST(WideInteger, DecimalFormatting)
+{
+    Decimal128 x(parse<Int128>("148873535527910577765226390751398592640"));
+
+    EXPECT_EQ(toString(x.value), "148873535527910577765226390751398592640");
+    EXPECT_EQ(toString(x.value / 10), "14887353552791057776522639075139859264");
+    EXPECT_EQ(toString(x.value % 10), "0");
+
+    Int128 fractional = DecimalUtils::getFractionalPart(x, 2);
+
+    EXPECT_EQ(fractional, 40);
+    EXPECT_EQ(decimalFractional(fractional, 2), "40");
+}

src/Compression/getCompressionCodecForFile.cpp

@@ -6,8 +6,6 @@
 #include <Common/PODArray.h>
 #include <common/logger_useful.h>
 
-#include <Common/UInt128.h>
-
 namespace DB
 {
 

src/Core/AccurateComparison.h

@@ -2,11 +2,11 @@
 
 #include <cmath>
 #include <limits>
-#include "Defines.h"
-#include "Types.h"
+#include <common/DecomposedFloat.h>
+#include <Core/Defines.h>
+#include <Core/Types.h>
 #include <common/extended_types.h>
 #include <Common/NaNUtils.h>
-#include <Common/UInt128.h>
 
 /** Preceptually-correct number comparisons.
   * Example: Int8(-1) != UInt8(255)
@@ -15,504 +15,142 @@
 namespace accurate
 {
 
-using DB::UInt64;
+using namespace DB;
 
-/** Cases:
-    1) Safe conversion (in case of default C++ operators)
-        a) int vs any int
-        b) uint vs any uint
-        c) float vs any float
-    2) int vs uint
-        a) sizeof(int) <= sizeof(uint). Accurate comparison with MAX_INT tresholds
-        b) sizeof(int)  > sizeof(uint). Casting to int
-    3) integral_type vs floating_type
-        a) sizeof(integral_type) <= 4. Comparison via casting arguments to Float64
-        b) sizeof(integral_type) == 8. Accurate comparison. Consider 3 sets of intervals:
-            1) interval between adjacent floats less or equal 1
-            2) interval between adjacent floats greater then 2
-            3) float is outside [MIN_INT64; MAX_INT64]
-*/
 
-// Case 1. Is pair of floats or pair of ints or pair of uints
 template <typename A, typename B>
-constexpr bool is_safe_conversion = (std::is_floating_point_v<A> && std::is_floating_point_v<B>)
-    || (is_integer_v<A> && is_integer_v<B> && !(is_signed_v<A> ^ is_signed_v<B>));
+bool lessOp(A a, B b)
+{
+    if constexpr (std::is_same_v<A, B>)
+        return a < b;
+
+    /// float vs float
+    if constexpr (std::is_floating_point_v<A> && std::is_floating_point_v<B>)
+        return a < b;
+
+    /// anything vs NaN
+    if (isNaN(a) || isNaN(b))
+        return false;
+
+    /// int vs int
+    if constexpr (is_integer_v<A> && is_integer_v<B>)
+    {
+        /// same signedness
+        if constexpr (is_signed_v<A> == is_signed_v<B>)
+            return a < b;
+
+        /// different signedness
+
+        if constexpr (is_signed_v<A> && !is_signed_v<B>)
+            return a < 0 || static_cast<make_unsigned_t<A>>(a) < b;
+
+        if constexpr (!is_signed_v<A> && is_signed_v<B>)
+            return b >= 0 && a < static_cast<make_unsigned_t<B>>(b);
+    }
+
+    /// int vs float
+    if constexpr (is_integer_v<A> && std::is_floating_point_v<B>)
+    {
+        if constexpr (sizeof(A) <= 4)
+            return static_cast<double>(a) < static_cast<double>(b);
+
+        return DecomposedFloat<B>(b).greater(a);
+    }
+
+    if constexpr (std::is_floating_point_v<A> && is_integer_v<B>)
+    {
+        if constexpr (sizeof(B) <= 4)
+            return static_cast<double>(a) < static_cast<double>(b);
+
+        return DecomposedFloat<A>(a).less(b);
+    }
+
+    static_assert(is_integer_v<A> || std::is_floating_point_v<A>);
+    static_assert(is_integer_v<B> || std::is_floating_point_v<B>);
+    __builtin_unreachable();
+}
+
 template <typename A, typename B>
-using bool_if_safe_conversion = std::enable_if_t<is_safe_conversion<A, B>, bool>;
+bool greaterOp(A a, B b)
+{
+    return lessOp(b, a);
+}
+
 template <typename A, typename B>
-using bool_if_not_safe_conversion = std::enable_if_t<!is_safe_conversion<A, B>, bool>;
-
-
-/// Case 2. Are params IntXX and UIntYY ?
-template <typename TInt, typename TUInt>
-constexpr bool is_any_int_vs_uint
-    = is_integer_v<TInt> && is_integer_v<TUInt> && is_signed_v<TInt> && is_unsigned_v<TUInt>;
-
-// Case 2a. Are params IntXX and UIntYY and sizeof(IntXX) >= sizeof(UIntYY) (in such case will use accurate compare)
-template <typename TInt, typename TUInt>
-constexpr bool is_le_int_vs_uint = is_any_int_vs_uint<TInt, TUInt> && (sizeof(TInt) <= sizeof(TUInt));
-
-static_assert(is_le_int_vs_uint<Int128, DB::UInt128>);
-static_assert(is_le_int_vs_uint<Int128, DB::UInt256>);
-
-template <typename TInt, typename TUInt>
-using bool_if_le_int_vs_uint_t = std::enable_if_t<is_le_int_vs_uint<TInt, TUInt>, bool>;
-
-template <typename TInt, typename TUInt>
-inline bool_if_le_int_vs_uint_t<TInt, TUInt> greaterOpTmpl(TInt a, TUInt b)
+bool greaterOrEqualsOp(A a, B b)
 {
-    return static_cast<TUInt>(a) > b && a >= 0 && b <= static_cast<TUInt>(std::numeric_limits<TInt>::max());
+    if (isNaN(a) || isNaN(b))
+        return false;
+
+    return !lessOp(a, b);
 }
 
-template <typename TUInt, typename TInt>
-inline bool_if_le_int_vs_uint_t<TInt, TUInt> greaterOpTmpl(TUInt a, TInt b)
+template <typename A, typename B>
+bool lessOrEqualsOp(A a, B b)
 {
-    return a > static_cast<TUInt>(b) || b < 0 || a > static_cast<TUInt>(std::numeric_limits<TInt>::max());
+    if (isNaN(a) || isNaN(b))
+        return false;
+
+    return !lessOp(b, a);
 }
 
-template <typename TInt, typename TUInt>
-inline bool_if_le_int_vs_uint_t<TInt, TUInt> equalsOpTmpl(TInt a, TUInt b)
+template <typename A, typename B>
+bool equalsOp(A a, B b)
 {
-    return static_cast<TUInt>(a) == b && a >= 0 && b <= static_cast<TUInt>(std::numeric_limits<TInt>::max());
-}
+    if constexpr (std::is_same_v<A, B>)
+        return a == b;
 
-template <typename TUInt, typename TInt>
-inline bool_if_le_int_vs_uint_t<TInt, TUInt> equalsOpTmpl(TUInt a, TInt b)
-{
-    return a == static_cast<TUInt>(b) && b >= 0 && a <= static_cast<TUInt>(std::numeric_limits<TInt>::max());
-}
+    /// float vs float
+    if constexpr (std::is_floating_point_v<A> && std::is_floating_point_v<B>)
+        return a == b;
 
+    /// anything vs NaN
+    if (isNaN(a) || isNaN(b))
+        return false;
 
-// Case 2b. Are params IntXX and UIntYY and sizeof(IntXX) > sizeof(UIntYY) (in such case will cast UIntYY to IntXX and compare)
-template <typename TInt, typename TUInt>
-constexpr bool is_gt_int_vs_uint = is_any_int_vs_uint<TInt, TUInt> && (sizeof(TInt) > sizeof(TUInt));
+    /// int vs int
+    if constexpr (is_integer_v<A> && is_integer_v<B>)
+    {
+        /// same signedness
+        if constexpr (is_signed_v<A> == is_signed_v<B>)
+            return a == b;
 
-template <typename TInt, typename TUInt>
-using bool_if_gt_int_vs_uint = std::enable_if_t<is_gt_int_vs_uint<TInt, TUInt>, bool>;
+        /// different signedness
 
-template <typename TInt, typename TUInt>
-inline bool_if_gt_int_vs_uint<TInt, TUInt> greaterOpTmpl(TInt a, TUInt b)
-{
-    return static_cast<TInt>(a) > static_cast<TInt>(b);
-}
+        if constexpr (is_signed_v<A> && !is_signed_v<B>)
+            return a >= 0 && static_cast<make_unsigned_t<A>>(a) == b;
 
-template <typename TInt, typename TUInt>
-inline bool_if_gt_int_vs_uint<TInt, TUInt> greaterOpTmpl(TUInt a, TInt b)
-{
-    using CastA = std::conditional_t<is_big_int_v<TInt> && std::is_same_v<TUInt, DB::UInt128>, DB::UInt256, TInt>;
+        if constexpr (!is_signed_v<A> && is_signed_v<B>)
+            return b >= 0 && a == static_cast<make_unsigned_t<B>>(b);
+    }
 
-    return static_cast<CastA>(a) > b;
-}
+    /// int vs float
+    if constexpr (is_integer_v<A> && std::is_floating_point_v<B>)
+    {
+        if constexpr (sizeof(A) <= 4)
+            return static_cast<double>(a) == static_cast<double>(b);
 
-template <typename TInt, typename TUInt>
-inline bool_if_gt_int_vs_uint<TInt, TUInt> equalsOpTmpl(TInt a, TUInt b)
-{
-    return static_cast<TInt>(a) == static_cast<TInt>(b);
-}
+        return DecomposedFloat<B>(b).equals(a);
+    }
 
-template <typename TInt, typename TUInt>
-inline bool_if_gt_int_vs_uint<TInt, TUInt> equalsOpTmpl(TUInt a, TInt b)
-{
-    return static_cast<TInt>(a) == static_cast<TInt>(b);
-}
+    if constexpr (std::is_floating_point_v<A> && is_integer_v<B>)
+    {
+        if constexpr (sizeof(B) <= 4)
+            return static_cast<double>(a) == static_cast<double>(b);
 
+        return DecomposedFloat<A>(a).equals(b);
+    }
 
-// Case 3a. Comparison via conversion to double.
-template <typename TAInt, typename TAFloat>
-using bool_if_double_can_be_used
-    = std::enable_if_t<is_integer_v<TAInt> && (sizeof(TAInt) <= 4) && std::is_floating_point_v<TAFloat>, bool>;
-
-template <typename TAInt, typename TAFloat>
-inline bool_if_double_can_be_used<TAInt, TAFloat> greaterOpTmpl(TAInt a, TAFloat b)
-{
-    return static_cast<double>(a) > static_cast<double>(b);
-}
-
-template <typename TAInt, typename TAFloat>
-inline bool_if_double_can_be_used<TAInt, TAFloat> greaterOpTmpl(TAFloat a, TAInt b)
-{
-    return static_cast<double>(a) > static_cast<double>(b);
-}
-
-template <typename TAInt, typename TAFloat>
-inline bool_if_double_can_be_used<TAInt, TAFloat> equalsOpTmpl(TAInt a, TAFloat b)
-{
-    return static_cast<double>(a) == static_cast<double>(b);
-}
-
-template <typename TAInt, typename TAFloat>
-inline bool_if_double_can_be_used<TAInt, TAFloat> equalsOpTmpl(TAFloat a, TAInt b)
-{
-    return static_cast<double>(a) == static_cast<double>(b);
-}
-
-// Big integers vs Float (not equal in any case for now, until big floats are introduced?)
-template <typename TABigInt, typename TAFloat>
-constexpr bool if_big_int_vs_float = (is_big_int_v<TABigInt> && std::is_floating_point_v<TAFloat>);
-
-template <typename TABigInt, typename TAFloat>
-using bool_if_big_int_vs_float = std::enable_if_t<if_big_int_vs_float<TABigInt, TAFloat>, bool>;
-
-template <typename TABigInt, typename TAFloat>
-inline bool_if_big_int_vs_float<TABigInt, TAFloat> greaterOpTmpl(TAFloat, TABigInt)
-{
+    /// e.g comparing UUID with integer.
     return false;
 }
 
-template <typename TABigInt, typename TAFloat>
-inline bool_if_big_int_vs_float<TABigInt, TAFloat> greaterOpTmpl(TABigInt, TAFloat)
-{
-    return false;
-}
-
-template <typename TABigInt, typename TAFloat>
-inline bool_if_big_int_vs_float<TABigInt, TAFloat> equalsOpTmpl(TAFloat, TABigInt)
-{
-    return false;
-}
-
-template <typename TABigInt, typename TAFloat>
-inline bool_if_big_int_vs_float<TABigInt, TAFloat> equalsOpTmpl(TABigInt, TAFloat)
-{
-    return false;
-}
-
-/* Final implementations */
-
-
 template <typename A, typename B>
-inline bool_if_not_safe_conversion<A, B> greaterOp(A a, B b)
-{
-    return greaterOpTmpl(a, b);
-}
-
-template <typename A, typename B>
-inline bool_if_safe_conversion<A, B> greaterOp(A a, B b)
-{
-    using CastA1 = std::conditional_t<is_big_int_v<B> && std::is_same_v<A, UInt8>, uint8_t, A>;
-    using CastB1 = std::conditional_t<is_big_int_v<A> && std::is_same_v<B, UInt8>, uint8_t, B>;
-
-    using CastA = std::conditional_t<is_big_int_v<B> && std::is_same_v<A, DB::UInt128>, B, CastA1>;
-    using CastB = std::conditional_t<is_big_int_v<A> && std::is_same_v<B, DB::UInt128>, A, CastB1>;
-
-    if constexpr (is_big_int_v<A> || is_big_int_v<B>)
-        return static_cast<CastA>(a) > static_cast<CastB>(b);
-    else
-        return a > b;
-}
-
-// Case 3b. 64-bit integers vs floats comparison.
-// See hint at https://github.com/JuliaLang/julia/issues/257 (but it doesn't work properly for -2**63)
-
-constexpr DB::Int64 MAX_INT64_WITH_EXACT_FLOAT64_REPR = 9007199254740992LL; // 2^53
-
-template <>
-inline bool greaterOp<DB::Float64, DB::Int64>(DB::Float64 f, DB::Int64 i)
-{
-    if (-MAX_INT64_WITH_EXACT_FLOAT64_REPR <= i && i <= MAX_INT64_WITH_EXACT_FLOAT64_REPR)
-        return f > static_cast<DB::Float64>(i);
-
-    return (f >= static_cast<DB::Float64>(std::numeric_limits<DB::Int64>::max())) // rhs is 2**63 (not 2^63 - 1)
-            || (f > static_cast<DB::Float64>(std::numeric_limits<DB::Int64>::min()) && static_cast<DB::Int64>(f) > i);
-}
-
-template <>
-inline bool greaterOp<DB::Int64, DB::Float64>(DB::Int64 i, DB::Float64 f)
-{
-    if (-MAX_INT64_WITH_EXACT_FLOAT64_REPR <= i && i <= MAX_INT64_WITH_EXACT_FLOAT64_REPR)
-        return f < static_cast<DB::Float64>(i);
-
-    return (f < static_cast<DB::Float64>(std::numeric_limits<DB::Int64>::min()))
-            || (f < static_cast<DB::Float64>(std::numeric_limits<DB::Int64>::max()) && i > static_cast<DB::Int64>(f));
-}
-
-template <>
-inline bool greaterOp<DB::Float64, DB::UInt64>(DB::Float64 f, DB::UInt64 u)
-{
-    if (u <= static_cast<DB::UInt64>(MAX_INT64_WITH_EXACT_FLOAT64_REPR))
-        return f > static_cast<DB::Float64>(u);
-
-    return (f >= static_cast<DB::Float64>(std::numeric_limits<DB::UInt64>::max()))
-            || (f >= 0 && static_cast<DB::UInt64>(f) > u);
-}
-
-template <>
-inline bool greaterOp<DB::UInt64, DB::Float64>(DB::UInt64 u, DB::Float64 f)
-{
-    if (u <= static_cast<DB::UInt64>(MAX_INT64_WITH_EXACT_FLOAT64_REPR))
-        return static_cast<DB::Float64>(u) > f;
-
-    return (f < 0)
-            || (f < static_cast<DB::Float64>(std::numeric_limits<DB::UInt64>::max()) && u > static_cast<UInt64>(f));
-}
-
-// Case 3b for float32
-template <>
-inline bool greaterOp<DB::Float32, DB::Int64>(DB::Float32 f, DB::Int64 i)
-{
-    return greaterOp(static_cast<DB::Float64>(f), i);
-}
-
-template <>
-inline bool greaterOp<DB::Int64, DB::Float32>(DB::Int64 i, DB::Float32 f)
-{
-    return greaterOp(i, static_cast<DB::Float64>(f));
-}
-
-template <>
-inline bool greaterOp<DB::Float32, DB::UInt64>(DB::Float32 f, DB::UInt64 u)
-{
-    return greaterOp(static_cast<DB::Float64>(f), u);
-}
-
-template <>
-inline bool greaterOp<DB::UInt64, DB::Float32>(DB::UInt64 u, DB::Float32 f)
-{
-    return greaterOp(u, static_cast<DB::Float64>(f));
-}
-
-template <>
-inline bool greaterOp<DB::Float64, DB::UInt128>(DB::Float64 f, DB::UInt128 u)
-{
-    /// TODO: This is wrong.
-    return u.low == 0 && greaterOp(f, u.high);
-}
-
-template <>
-inline bool greaterOp<DB::UInt128, DB::Float64>(DB::UInt128 u, DB::Float64 f)
-{
-    /// TODO: This is wrong.
-    return u.low != 0 || greaterOp(u.high, f);
-}
-
-template <>
-inline bool greaterOp<DB::Float32, DB::UInt128>(DB::Float32 f, DB::UInt128 u)
-{
-    return greaterOp(static_cast<DB::Float64>(f), u);
-}
-
-template <>
-inline bool greaterOp<DB::UInt128, DB::Float32>(DB::UInt128 u, DB::Float32 f)
-{
-    return greaterOp(u, static_cast<DB::Float64>(f));
-}
-
-template <typename A, typename B>
-inline bool_if_not_safe_conversion<A, B> equalsOp(A a, B b)
-{
-    return equalsOpTmpl(a, b);
-}
-
-template <typename A, typename B>
-inline bool_if_safe_conversion<A, B> equalsOp(A a, B b)
-{
-    using LargestType = std::conditional_t<(sizeof(A) > sizeof(B)) || ((sizeof(A) == sizeof(B)) && !std::is_same_v<A, DB::UInt128>), A, B>;
-
-    return static_cast<LargestType>(a) == static_cast<LargestType>(b);
-}
-
-template <>
-inline bool equalsOp<DB::Float64, DB::UInt64>(DB::Float64 f, DB::UInt64 u)
-{
-    /// Maximum exactly representable integer.
-    return u <= (1ULL << std::numeric_limits<DB::Float64>::digits)
-        && f == static_cast<DB::Float64>(u);
-}
-
-template <>
-inline bool equalsOp<DB::UInt64, DB::Float64>(DB::UInt64 u, DB::Float64 f)
-{
-    return equalsOp(f, u);
-}
-
-template <>
-inline bool equalsOp<DB::Float64, DB::Int64>(DB::Float64 f, DB::Int64 u)
-{
-    return u <= (1LL << std::numeric_limits<DB::Float64>::digits)
-        && u >= -(1LL << std::numeric_limits<DB::Float64>::digits)
-        && f == static_cast<DB::Float64>(u);
-}
-
-template <>
-inline bool equalsOp<DB::Int64, DB::Float64>(DB::Int64 u, DB::Float64 f)
-{
-    return equalsOp(f, u);
-}
-
-template <>
-inline bool equalsOp<DB::Float32, DB::UInt64>(DB::Float32 f, DB::UInt64 u)
-{
-    return u <= (1ULL << std::numeric_limits<DB::Float32>::digits)
-        && f == static_cast<DB::Float32>(u);
-}
-
-template <>
-inline bool equalsOp<DB::UInt64, DB::Float32>(DB::UInt64 u, DB::Float32 f)
-{
-    return equalsOp(f, u);
-}
-
-template <>
-inline bool equalsOp<DB::Float32, DB::Int64>(DB::Float32 f, DB::Int64 u)
-{
-    return u <= (1LL << std::numeric_limits<DB::Float32>::digits)
-        && u >= -(1LL << std::numeric_limits<DB::Float32>::digits)
-        && f == static_cast<DB::Float32>(u);
-}
-
-template <>
-inline bool equalsOp<DB::Int64, DB::Float32>(DB::Int64 u, DB::Float32 f)
-{
-    return equalsOp(f, u);
-}
-
-template <>
-inline bool equalsOp<DB::UInt128, DB::Float64>(DB::UInt128 u, DB::Float64 f)
-{
-    /// TODO: This is wrong.
-    return u.low == 0 && equalsOp(static_cast<UInt64>(u.high), f);
-}
-
-template <>
-inline bool equalsOp<DB::UInt128, DB::Float32>(DB::UInt128 u, DB::Float32 f)
-{
-    return equalsOp(u, static_cast<DB::Float64>(f));
-}
-
-template <>
-inline bool equalsOp<DB::Float64, DB::UInt128>(DB::Float64 f, DB::UInt128 u)
-{
-    return equalsOp(u, f);
-}
-
-template <>
-inline bool equalsOp<DB::Float32, DB::UInt128>(DB::Float32 f, DB::UInt128 u)
-{
-    return equalsOp(static_cast<DB::Float64>(f), u);
-}
-
-inline bool NO_SANITIZE_UNDEFINED greaterOp(DB::Int128 i, DB::Float64 f)
-{
-    static constexpr Int128 min_int128 = minInt128();
-    static constexpr Int128 max_int128 = maxInt128();
-
-    if (-MAX_INT64_WITH_EXACT_FLOAT64_REPR <= i && i <= MAX_INT64_WITH_EXACT_FLOAT64_REPR)
-        return static_cast<DB::Float64>(i) > f;
-
-    return (f < static_cast<DB::Float64>(min_int128))
-        || (f < static_cast<DB::Float64>(max_int128) && i > static_cast<DB::Int128>(f));
-}
-
-inline bool NO_SANITIZE_UNDEFINED greaterOp(DB::Float64 f, DB::Int128 i)
-{
-    static constexpr Int128 min_int128 = minInt128();
-    static constexpr Int128 max_int128 = maxInt128();
-
-    if (-MAX_INT64_WITH_EXACT_FLOAT64_REPR <= i && i <= MAX_INT64_WITH_EXACT_FLOAT64_REPR)
-        return f > static_cast<DB::Float64>(i);
-
-    return (f >= static_cast<DB::Float64>(max_int128))
-        || (f > static_cast<DB::Float64>(min_int128) && static_cast<DB::Int128>(f) > i);
-}
-
-inline bool greaterOp(DB::Int128 i, DB::Float32 f) { return greaterOp(i, static_cast<DB::Float64>(f)); }
-inline bool greaterOp(DB::Float32 f, DB::Int128 i) { return greaterOp(static_cast<DB::Float64>(f), i); }
-
-inline bool NO_SANITIZE_UNDEFINED equalsOp(DB::Int128 i, DB::Float64 f) { return i == static_cast<DB::Int128>(f) && static_cast<DB::Float64>(i) == f; }
-inline bool NO_SANITIZE_UNDEFINED equalsOp(DB::Int128 i, DB::Float32 f) { return i == static_cast<DB::Int128>(f) && static_cast<DB::Float32>(i) == f; }
-inline bool equalsOp(DB::Float64 f, DB::Int128 i) { return equalsOp(i, f); }
-inline bool equalsOp(DB::Float32 f, DB::Int128 i) { return equalsOp(i, f); }
-
-template <typename A, typename B>
-inline bool_if_not_safe_conversion<A, B> notEqualsOp(A a, B b)
+bool notEqualsOp(A a, B b)
 {
     return !equalsOp(a, b);
 }
 
-template <typename A, typename B>
-inline bool_if_safe_conversion<A, B> notEqualsOp(A a, B b)
-{
-    using CastA1 = std::conditional_t<is_big_int_v<B> && std::is_same_v<A, UInt8>, uint8_t, A>;
-    using CastB1 = std::conditional_t<is_big_int_v<A> && std::is_same_v<B, UInt8>, uint8_t, B>;
-
-    using CastA = std::conditional_t<is_big_int_v<B> && std::is_same_v<A, DB::UInt128>, B, CastA1>;
-    using CastB = std::conditional_t<is_big_int_v<A> && std::is_same_v<B, DB::UInt128>, A, CastB1>;
-
-    if constexpr (is_big_int_v<A> || is_big_int_v<B>)
-        return static_cast<CastA>(a) != static_cast<CastB>(b);
-    else
-        return a != b;
-}
-
-
-template <typename A, typename B>
-inline bool_if_not_safe_conversion<A, B> lessOp(A a, B b)
-{
-    return greaterOp(b, a);
-}
-
-template <typename A, typename B>
-inline bool_if_safe_conversion<A, B> lessOp(A a, B b)
-{
-    using CastA1 = std::conditional_t<is_big_int_v<B> && std::is_same_v<A, UInt8>, uint8_t, A>;
-    using CastB1 = std::conditional_t<is_big_int_v<A> && std::is_same_v<B, UInt8>, uint8_t, B>;
-
-    using CastA = std::conditional_t<is_big_int_v<B> && std::is_same_v<A, DB::UInt128>, B, CastA1>;
-    using CastB = std::conditional_t<is_big_int_v<A> && std::is_same_v<B, DB::UInt128>, A, CastB1>;
-
-    if constexpr (is_big_int_v<A> || is_big_int_v<B>)
-        return static_cast<CastA>(a) < static_cast<CastB>(b);
-    else
-        return a < b;
-}
-
-
-template <typename A, typename B>
-inline bool_if_not_safe_conversion<A, B> lessOrEqualsOp(A a, B b)
-{
-    if (isNaN(a) || isNaN(b))
-        return false;
-    return !greaterOp(a, b);
-}
-
-template <typename A, typename B>
-inline bool_if_safe_conversion<A, B> lessOrEqualsOp(A a, B b)
-{
-    using CastA1 = std::conditional_t<is_big_int_v<B> && std::is_same_v<A, UInt8>, uint8_t, A>;
-    using CastB1 = std::conditional_t<is_big_int_v<A> && std::is_same_v<B, UInt8>, uint8_t, B>;
-
-    using CastA = std::conditional_t<is_big_int_v<B> && std::is_same_v<A, DB::UInt128>, B, CastA1>;
-    using CastB = std::conditional_t<is_big_int_v<A> && std::is_same_v<B, DB::UInt128>, A, CastB1>;
-
-    if constexpr (is_big_int_v<A> || is_big_int_v<B>)
-        return static_cast<CastA>(a) <= static_cast<CastB>(b);
-    else
-        return a <= b;
-}
-
-
-template <typename A, typename B>
-inline bool_if_not_safe_conversion<A, B> greaterOrEqualsOp(A a, B b)
-{
-    if (isNaN(a) || isNaN(b))
-        return false;
-    return !greaterOp(b, a);
-}
-
-template <typename A, typename B>
-inline bool_if_safe_conversion<A, B> greaterOrEqualsOp(A a, B b)
-{
-    using CastA1 = std::conditional_t<is_big_int_v<B> && std::is_same_v<A, UInt8>, uint8_t, A>;
-    using CastB1 = std::conditional_t<is_big_int_v<A> && std::is_same_v<B, UInt8>, uint8_t, B>;
-
-    using CastA = std::conditional_t<is_big_int_v<B> && std::is_same_v<A, DB::UInt128>, B, CastA1>;
-    using CastB = std::conditional_t<is_big_int_v<A> && std::is_same_v<B, DB::UInt128>, A, CastB1>;
-
-    if constexpr (is_big_int_v<A> || is_big_int_v<B>)
-        return static_cast<CastA>(a) >= static_cast<CastB>(b);
-    else
-        return a >= b;
-}
 
 /// Converts numeric to an equal numeric of other type.
 template <typename From, typename To>
@@ -547,8 +185,8 @@ inline bool NO_SANITIZE_UNDEFINED convertNumeric(From value, To & result)
         }
     }
 
-    if (accurate::greaterOp(value, std::numeric_limits<To>::max())
-        || accurate::greaterOp(std::numeric_limits<To>::lowest(), value))
+    if (greaterOp(value, std::numeric_limits<To>::max())
+        || lessOp(value, std::numeric_limits<To>::lowest()))
     {
         return false;
     }

src/Core/BaseSettings.h

@@ -2,7 +2,6 @@
 
 #include <Core/SettingsFields.h>
 #include <Common/SettingsChanges.h>
-#include <Common/FieldVisitors.h>
 #include <ext/range.h>
 #include <boost/blank.hpp>
 #include <unordered_map>

src/Core/DecimalComparison.h

@@ -36,10 +36,13 @@ inline bool allowDecimalComparison(const DataTypePtr & left_type, const DataType
     return false;
 }
 
-template <size_t > struct ConstructDecInt { using Type = Int32; };
+template <size_t> struct ConstructDecInt;
+template <> struct ConstructDecInt<1> { using Type = Int32; };
+template <> struct ConstructDecInt<2> { using Type = Int32; };
+template <> struct ConstructDecInt<4> { using Type = Int32; };
 template <> struct ConstructDecInt<8> { using Type = Int64; };
 template <> struct ConstructDecInt<16> { using Type = Int128; };
-template <> struct ConstructDecInt<48> { using Type = Int256; };
+template <> struct ConstructDecInt<32> { using Type = Int256; };
 
 template <typename T, typename U>
 struct DecCompareInt
@@ -249,7 +252,7 @@ private:
                 overflow |= common::mulOverflow(y, scale, y);
 
             if (overflow)
-                throw Exception("Can't compare", ErrorCodes::DECIMAL_OVERFLOW);
+                throw Exception("Can't compare decimal number due to overflow", ErrorCodes::DECIMAL_OVERFLOW);
         }
         else
         {

src/Core/DecimalFunctions.h

@@ -276,9 +276,7 @@ template <typename To, typename DecimalType>
 To convertTo(const DecimalType & decimal, size_t scale)
 {
     To result;
-
     convertToImpl<To, DecimalType, void>(decimal, scale, result);
-
     return result;
 }
 

src/Core/Field.cpp

@@ -238,6 +238,13 @@ Field Field::restoreFromDump(const std::string_view & dump_)
         return value;
     }
 
+    prefix = std::string_view{"UInt128_"};
+    if (dump.starts_with(prefix))
+    {
+        UInt128 value = parseFromString<UInt128>(dump.substr(prefix.length()));
+        return value;
+    }
+
     prefix = std::string_view{"Int256_"};
     if (dump.starts_with(prefix))
     {
@@ -295,15 +302,6 @@ Field Field::restoreFromDump(const std::string_view & dump_)
         return decimal;
     }
 
-    prefix = std::string_view{"UUID_"};
-    if (dump.starts_with(prefix))
-    {
-        UUID uuid;
-        ReadBufferFromString buf{dump.substr(prefix.length())};
-        readQuoted(uuid, buf);
-        return uuid;
-    }
-
     if (dump.starts_with("\'"))
     {
         String str;
@@ -410,45 +408,45 @@ Field Field::restoreFromDump(const std::string_view & dump_)
 
 
 template <typename T>
-static bool decEqual(T x, T y, UInt32 x_scale, UInt32 y_scale)
+bool decimalEqual(T x, T y, UInt32 x_scale, UInt32 y_scale)
 {
     using Comparator = DecimalComparison<T, T, EqualsOp>;
     return Comparator::compare(x, y, x_scale, y_scale);
 }
 
 template <typename T>
-static bool decLess(T x, T y, UInt32 x_scale, UInt32 y_scale)
+bool decimalLess(T x, T y, UInt32 x_scale, UInt32 y_scale)
 {
     using Comparator = DecimalComparison<T, T, LessOp>;
     return Comparator::compare(x, y, x_scale, y_scale);
 }
 
 template <typename T>
-static bool decLessOrEqual(T x, T y, UInt32 x_scale, UInt32 y_scale)
+bool decimalLessOrEqual(T x, T y, UInt32 x_scale, UInt32 y_scale)
 {
     using Comparator = DecimalComparison<T, T, LessOrEqualsOp>;
     return Comparator::compare(x, y, x_scale, y_scale);
 }
 
-template <> bool decimalEqual(Decimal32 x, Decimal32 y, UInt32 x_scale, UInt32 y_scale) { return decEqual(x, y, x_scale, y_scale); }
-template <> bool decimalLess(Decimal32 x, Decimal32 y, UInt32 x_scale, UInt32 y_scale) { return decLess(x, y, x_scale, y_scale); }
-template <> bool decimalLessOrEqual(Decimal32 x, Decimal32 y, UInt32 x_scale, UInt32 y_scale) { return decLessOrEqual(x, y, x_scale, y_scale); }
 
-template <> bool decimalEqual(Decimal64 x, Decimal64 y, UInt32 x_scale, UInt32 y_scale) { return decEqual(x, y, x_scale, y_scale); }
-template <> bool decimalLess(Decimal64 x, Decimal64 y, UInt32 x_scale, UInt32 y_scale) { return decLess(x, y, x_scale, y_scale); }
-template <> bool decimalLessOrEqual(Decimal64 x, Decimal64 y, UInt32 x_scale, UInt32 y_scale) { return decLessOrEqual(x, y, x_scale, y_scale); }
+template bool decimalEqual<Decimal32>(Decimal32 x, Decimal32 y, UInt32 x_scale, UInt32 y_scale);
+template bool decimalEqual<Decimal64>(Decimal64 x, Decimal64 y, UInt32 x_scale, UInt32 y_scale);
+template bool decimalEqual<Decimal128>(Decimal128 x, Decimal128 y, UInt32 x_scale, UInt32 y_scale);
+template bool decimalEqual<Decimal256>(Decimal256 x, Decimal256 y, UInt32 x_scale, UInt32 y_scale);
+template bool decimalEqual<DateTime64>(DateTime64 x, DateTime64 y, UInt32 x_scale, UInt32 y_scale);
 
-template <> bool decimalEqual(Decimal128 x, Decimal128 y, UInt32 x_scale, UInt32 y_scale) { return decEqual(x, y, x_scale, y_scale); }
-template <> bool decimalLess(Decimal128 x, Decimal128 y, UInt32 x_scale, UInt32 y_scale) { return decLess(x, y, x_scale, y_scale); }
-template <> bool decimalLessOrEqual(Decimal128 x, Decimal128 y, UInt32 x_scale, UInt32 y_scale) { return decLessOrEqual(x, y, x_scale, y_scale); }
+template bool decimalLess<Decimal32>(Decimal32 x, Decimal32 y, UInt32 x_scale, UInt32 y_scale);
+template bool decimalLess<Decimal64>(Decimal64 x, Decimal64 y, UInt32 x_scale, UInt32 y_scale);
+template bool decimalLess<Decimal128>(Decimal128 x, Decimal128 y, UInt32 x_scale, UInt32 y_scale);
+template bool decimalLess<Decimal256>(Decimal256 x, Decimal256 y, UInt32 x_scale, UInt32 y_scale);
+template bool decimalLess<DateTime64>(DateTime64 x, DateTime64 y, UInt32 x_scale, UInt32 y_scale);
 
-template <> bool decimalEqual(Decimal256 x, Decimal256 y, UInt32 x_scale, UInt32 y_scale) { return decEqual(x, y, x_scale, y_scale); }
-template <> bool decimalLess(Decimal256 x, Decimal256 y, UInt32 x_scale, UInt32 y_scale) { return decLess(x, y, x_scale, y_scale); }
-template <> bool decimalLessOrEqual(Decimal256 x, Decimal256 y, UInt32 x_scale, UInt32 y_scale) { return decLessOrEqual(x, y, x_scale, y_scale); }
+template bool decimalLessOrEqual<Decimal32>(Decimal32 x, Decimal32 y, UInt32 x_scale, UInt32 y_scale);
+template bool decimalLessOrEqual<Decimal64>(Decimal64 x, Decimal64 y, UInt32 x_scale, UInt32 y_scale);
+template bool decimalLessOrEqual<Decimal128>(Decimal128 x, Decimal128 y, UInt32 x_scale, UInt32 y_scale);
+template bool decimalLessOrEqual<Decimal256>(Decimal256 x, Decimal256 y, UInt32 x_scale, UInt32 y_scale);
+template bool decimalLessOrEqual<DateTime64>(DateTime64 x, DateTime64 y, UInt32 x_scale, UInt32 y_scale);
 
-template <> bool decimalEqual(DateTime64 x, DateTime64 y, UInt32 x_scale, UInt32 y_scale) { return decEqual(x, y, x_scale, y_scale); }
-template <> bool decimalLess(DateTime64 x, DateTime64 y, UInt32 x_scale, UInt32 y_scale) { return decLess(x, y, x_scale, y_scale); }
-template <> bool decimalLessOrEqual(DateTime64 x, DateTime64 y, UInt32 x_scale, UInt32 y_scale) { return decLessOrEqual(x, y, x_scale, y_scale); }
 
 inline void writeText(const Null &, WriteBuffer & buf)
 {

src/Core/Field.h

@@ -7,7 +7,6 @@
 #include <functional>
 
 #include <Common/Exception.h>
-#include <Common/UInt128.h>
 #include <Common/AllocatorWithMemoryTracking.h>
 #include <Core/Types.h>
 #include <Core/Defines.h>
@@ -177,8 +176,7 @@ template <> struct NearestFieldTypeImpl<UInt16> { using Type = UInt64; };
 template <> struct NearestFieldTypeImpl<UInt32> { using Type = UInt64; };
 
 template <> struct NearestFieldTypeImpl<DayNum> { using Type = UInt64; };
-template <> struct NearestFieldTypeImpl<UInt128> { using Type = UInt128; };
-template <> struct NearestFieldTypeImpl<UUID> { using Type = UInt128; };
+template <> struct NearestFieldTypeImpl<UUID> { using Type = UUID; };
 template <> struct NearestFieldTypeImpl<Int16> { using Type = Int64; };
 template <> struct NearestFieldTypeImpl<Int32> { using Type = Int64; };
 
@@ -191,8 +189,9 @@ template <> struct NearestFieldTypeImpl<unsigned long long> { using Type = UInt6
 
 template <> struct NearestFieldTypeImpl<UInt256> { using Type = UInt256; };
 template <> struct NearestFieldTypeImpl<Int256> { using Type = Int256; };
-
+template <> struct NearestFieldTypeImpl<UInt128> { using Type = UInt128; };
 template <> struct NearestFieldTypeImpl<Int128> { using Type = Int128; };
+
 template <> struct NearestFieldTypeImpl<Decimal32> { using Type = DecimalField<Decimal32>; };
 template <> struct NearestFieldTypeImpl<Decimal64> { using Type = DecimalField<Decimal64>; };
 template <> struct NearestFieldTypeImpl<Decimal128> { using Type = DecimalField<Decimal128>; };
@@ -235,7 +234,7 @@ struct NearestFieldTypeImpl<T, std::enable_if_t<std::is_enum_v<T>>>
   *  but somewhat more efficient, and simpler.
   *
   * Used to represent a single value of one of several types in memory.
-  * Warning! Prefer to use chunks of columns instead of single values. See Column.h
+  * Warning! Prefer to use chunks of columns instead of single values. See IColumn.h
   */
 class Field
 {
@@ -252,8 +251,6 @@ public:
             UInt128 = 4,
             Int128  = 5,
 
-            /// Non-POD types.
-
             String  = 16,
             Array   = 17,
             Tuple   = 18,
@@ -265,10 +262,9 @@ public:
             UInt256 = 24,
             Int256  = 25,
             Map = 26,
+            UUID = 27,
         };
 
-        static const int MIN_NON_POD = 16;
-
         static const char * toString(Which which)
         {
             switch (which)
@@ -276,8 +272,11 @@ public:
                 case Null:    return "Null";
                 case UInt64:  return "UInt64";
                 case UInt128: return "UInt128";
+                case UInt256: return "UInt256";
                 case Int64:   return "Int64";
                 case Int128:  return "Int128";
+                case Int256:  return "Int256";
+                case UUID:    return "UUID";
                 case Float64: return "Float64";
                 case String:  return "String";
                 case Array:   return "Array";
@@ -288,8 +287,6 @@ public:
                 case Decimal128: return "Decimal128";
                 case Decimal256: return "Decimal256";
                 case AggregateFunctionState: return "AggregateFunctionState";
-                case UInt256: return "UInt256";
-                case Int256:  return "Int256";
             }
 
             throw Exception("Bad type of Field", ErrorCodes::BAD_TYPE_OF_FIELD);
@@ -303,12 +300,17 @@ public:
 
     static bool IsDecimal(Types::Which which)
     {
-        return (which >= Types::Decimal32 && which <= Types::Decimal128) || which == Types::Decimal256;
+        return which == Types::Decimal32
+            || which == Types::Decimal64
+            || which == Types::Decimal128
+            || which == Types::Decimal256;
     }
 
     /// Templates to avoid ambiguity.
     template <typename T, typename Z = void *>
-    using enable_if_not_field_or_stringlike_t = std::enable_if_t<!std::is_same_v<std::decay_t<T>, Field> && !std::is_same_v<NearestFieldType<std::decay_t<T>>, String>, Z>;
+    using enable_if_not_field_or_stringlike_t = std::enable_if_t<
+        !std::is_same_v<std::decay_t<T>, Field>
+        && !std::is_same_v<NearestFieldType<std::decay_t<T>>, String>, Z>;
 
     Field() //-V730
         : which(Types::Null)
@@ -382,10 +384,10 @@ public:
     enable_if_not_field_or_stringlike_t<T, Field> &
     operator=(T && rhs);
 
-    Field & operator =(const std::string_view & str);
-    Field & operator =(const String & str) { return *this = std::string_view{str}; }
-    Field & operator =(String && str);
-    Field & operator =(const char * str) { return *this = std::string_view{str}; }
+    Field & operator= (const std::string_view & str);
+    Field & operator= (const String & str) { return *this = std::string_view{str}; }
+    Field & operator= (String && str);
+    Field & operator= (const char * str) { return *this = std::string_view{str}; }
 
     ~Field()
     {
@@ -454,8 +456,11 @@ public:
             case Types::Null:    return false;
             case Types::UInt64:  return get<UInt64>()  < rhs.get<UInt64>();
             case Types::UInt128: return get<UInt128>() < rhs.get<UInt128>();
+            case Types::UInt256: return get<UInt256>() < rhs.get<UInt256>();
             case Types::Int64:   return get<Int64>()   < rhs.get<Int64>();
             case Types::Int128:  return get<Int128>()  < rhs.get<Int128>();
+            case Types::Int256:  return get<Int256>()  < rhs.get<Int256>();
+            case Types::UUID:    return get<UUID>()    < rhs.get<UUID>();
             case Types::Float64: return get<Float64>() < rhs.get<Float64>();
             case Types::String:  return get<String>()  < rhs.get<String>();
             case Types::Array:   return get<Array>()   < rhs.get<Array>();
@@ -466,8 +471,6 @@ public:
             case Types::Decimal128: return get<DecimalField<Decimal128>>() < rhs.get<DecimalField<Decimal128>>();
             case Types::Decimal256: return get<DecimalField<Decimal256>>() < rhs.get<DecimalField<Decimal256>>();
             case Types::AggregateFunctionState:  return get<AggregateFunctionStateData>() < rhs.get<AggregateFunctionStateData>();
-            case Types::UInt256: return get<UInt256>() < rhs.get<UInt256>();
-            case Types::Int256: return get<Int256>() < rhs.get<Int256>();
         }
 
         throw Exception("Bad type of Field", ErrorCodes::BAD_TYPE_OF_FIELD);
@@ -490,8 +493,11 @@ public:
             case Types::Null:    return true;
             case Types::UInt64:  return get<UInt64>()  <= rhs.get<UInt64>();
             case Types::UInt128: return get<UInt128>() <= rhs.get<UInt128>();
+            case Types::UInt256: return get<UInt256>() <= rhs.get<UInt256>();
             case Types::Int64:   return get<Int64>()   <= rhs.get<Int64>();
             case Types::Int128:  return get<Int128>()  <= rhs.get<Int128>();
+            case Types::Int256:  return get<Int256>()  <= rhs.get<Int256>();
+            case Types::UUID:    return get<UUID>().toUnderType() <= rhs.get<UUID>().toUnderType();
             case Types::Float64: return get<Float64>() <= rhs.get<Float64>();
             case Types::String:  return get<String>()  <= rhs.get<String>();
             case Types::Array:   return get<Array>()   <= rhs.get<Array>();
@@ -502,8 +508,6 @@ public:
             case Types::Decimal128: return get<DecimalField<Decimal128>>() <= rhs.get<DecimalField<Decimal128>>();
             case Types::Decimal256: return get<DecimalField<Decimal256>>() <= rhs.get<DecimalField<Decimal256>>();
             case Types::AggregateFunctionState:  return get<AggregateFunctionStateData>() <= rhs.get<AggregateFunctionStateData>();
-            case Types::UInt256: return get<UInt256>() <= rhs.get<UInt256>();
-            case Types::Int256: return get<Int256>() <= rhs.get<Int256>();
         }
 
         throw Exception("Bad type of Field", ErrorCodes::BAD_TYPE_OF_FIELD);
@@ -531,19 +535,20 @@ public:
                 // Compare as UInt64 so that NaNs compare as equal.
                 return reinterpret<UInt64>() == rhs.reinterpret<UInt64>();
             }
+            case Types::UUID:    return get<UUID>()    == rhs.get<UUID>();
             case Types::String:  return get<String>()  == rhs.get<String>();
             case Types::Array:   return get<Array>()   == rhs.get<Array>();
             case Types::Tuple:   return get<Tuple>()   == rhs.get<Tuple>();
             case Types::Map:     return get<Map>()     == rhs.get<Map>();
             case Types::UInt128: return get<UInt128>() == rhs.get<UInt128>();
+            case Types::UInt256: return get<UInt256>() == rhs.get<UInt256>();
             case Types::Int128:  return get<Int128>()  == rhs.get<Int128>();
+            case Types::Int256:  return get<Int256>()  == rhs.get<Int256>();
             case Types::Decimal32:  return get<DecimalField<Decimal32>>()  == rhs.get<DecimalField<Decimal32>>();
             case Types::Decimal64:  return get<DecimalField<Decimal64>>()  == rhs.get<DecimalField<Decimal64>>();
             case Types::Decimal128: return get<DecimalField<Decimal128>>() == rhs.get<DecimalField<Decimal128>>();
             case Types::Decimal256: return get<DecimalField<Decimal256>>() == rhs.get<DecimalField<Decimal256>>();
             case Types::AggregateFunctionState:  return get<AggregateFunctionStateData>() == rhs.get<AggregateFunctionStateData>();
-            case Types::UInt256: return get<UInt256>() == rhs.get<UInt256>();
-            case Types::Int256:  return get<Int256>()  == rhs.get<Int256>();
         }
 
         throw Exception("Bad type of Field", ErrorCodes::BAD_TYPE_OF_FIELD);
@@ -569,7 +574,11 @@ public:
 #endif
             case Types::UInt64:  return f(field.template get<UInt64>());
             case Types::UInt128: return f(field.template get<UInt128>());
+            case Types::UInt256: return f(field.template get<UInt256>());
             case Types::Int64:   return f(field.template get<Int64>());
+            case Types::Int128:  return f(field.template get<Int128>());
+            case Types::Int256:  return f(field.template get<Int256>());
+            case Types::UUID:    return f(field.template get<UUID>());
             case Types::Float64: return f(field.template get<Float64>());
             case Types::String:  return f(field.template get<String>());
             case Types::Array:   return f(field.template get<Array>());
@@ -580,19 +589,12 @@ public:
             case Types::Decimal128: return f(field.template get<DecimalField<Decimal128>>());
             case Types::Decimal256: return f(field.template get<DecimalField<Decimal256>>());
             case Types::AggregateFunctionState: return f(field.template get<AggregateFunctionStateData>());
-            case Types::Int128: return f(field.template get<Int128>());
-            case Types::UInt256: return f(field.template get<UInt256>());
-            case Types::Int256: return f(field.template get<Int256>());
 #if !defined(__clang__)
 #pragma GCC diagnostic pop
 #endif
         }
 
-        // GCC 9 complains that control reaches the end, despite that we handle
-        // all the cases above (maybe because of throw?). Return something to
-        // silence it.
-        Null null{};
-        return f(null);
+        __builtin_unreachable();
     }
 
     String dump() const;
@@ -600,10 +602,9 @@ public:
 
 private:
     std::aligned_union_t<DBMS_MIN_FIELD_SIZE - sizeof(Types::Which),
-        Null, UInt64, UInt128, Int64, Int128, Float64, String, Array, Tuple, Map,
+        Null, UInt64, UInt128, UInt256, Int64, Int128, Int256, UUID, Float64, String, Array, Tuple, Map,
         DecimalField<Decimal32>, DecimalField<Decimal64>, DecimalField<Decimal128>, DecimalField<Decimal256>,
-        AggregateFunctionStateData,
-        UInt256, Int256
+        AggregateFunctionStateData
         > storage;
 
     Types::Which which;
@@ -685,9 +686,6 @@ private:
 
     ALWAYS_INLINE void destroy()
     {
-        if (which < Types::MIN_NON_POD)
-            return;
-
         switch (which)
         {
             case Types::String:
@@ -726,8 +724,11 @@ private:
 template <> struct Field::TypeToEnum<Null>    { static const Types::Which value = Types::Null; };
 template <> struct Field::TypeToEnum<UInt64>  { static const Types::Which value = Types::UInt64; };
 template <> struct Field::TypeToEnum<UInt128> { static const Types::Which value = Types::UInt128; };
+template <> struct Field::TypeToEnum<UInt256> { static const Types::Which value = Types::UInt256; };
 template <> struct Field::TypeToEnum<Int64>   { static const Types::Which value = Types::Int64; };
 template <> struct Field::TypeToEnum<Int128>  { static const Types::Which value = Types::Int128; };
+template <> struct Field::TypeToEnum<Int256>  { static const Types::Which value = Types::Int256; };
+template <> struct Field::TypeToEnum<UUID>    { static const Types::Which value = Types::UUID; };
 template <> struct Field::TypeToEnum<Float64> { static const Types::Which value = Types::Float64; };
 template <> struct Field::TypeToEnum<String>  { static const Types::Which value = Types::String; };
 template <> struct Field::TypeToEnum<Array>   { static const Types::Which value = Types::Array; };
@@ -739,14 +740,15 @@ template <> struct Field::TypeToEnum<DecimalField<Decimal128>>{ static const Typ
 template <> struct Field::TypeToEnum<DecimalField<Decimal256>>{ static const Types::Which value = Types::Decimal256; };
 template <> struct Field::TypeToEnum<DecimalField<DateTime64>>{ static const Types::Which value = Types::Decimal64; };
 template <> struct Field::TypeToEnum<AggregateFunctionStateData>{ static const Types::Which value = Types::AggregateFunctionState; };
-template <> struct Field::TypeToEnum<UInt256> { static const Types::Which value = Types::UInt256; };
-template <> struct Field::TypeToEnum<Int256> { static const Types::Which value = Types::Int256; };
 
 template <> struct Field::EnumToType<Field::Types::Null>    { using Type = Null; };
 template <> struct Field::EnumToType<Field::Types::UInt64>  { using Type = UInt64; };
 template <> struct Field::EnumToType<Field::Types::UInt128> { using Type = UInt128; };
+template <> struct Field::EnumToType<Field::Types::UInt256> { using Type = UInt256; };
 template <> struct Field::EnumToType<Field::Types::Int64>   { using Type = Int64; };
 template <> struct Field::EnumToType<Field::Types::Int128>  { using Type = Int128; };
+template <> struct Field::EnumToType<Field::Types::Int256>  { using Type = Int256; };
+template <> struct Field::EnumToType<Field::Types::UUID>    { using Type = UUID; };
 template <> struct Field::EnumToType<Field::Types::Float64> { using Type = Float64; };
 template <> struct Field::EnumToType<Field::Types::String>  { using Type = String; };
 template <> struct Field::EnumToType<Field::Types::Array>   { using Type = Array; };
@@ -757,10 +759,8 @@ template <> struct Field::EnumToType<Field::Types::Decimal64> { using Type = Dec
 template <> struct Field::EnumToType<Field::Types::Decimal128> { using Type = DecimalField<Decimal128>; };
 template <> struct Field::EnumToType<Field::Types::Decimal256> { using Type = DecimalField<Decimal256>; };
 template <> struct Field::EnumToType<Field::Types::AggregateFunctionState> { using Type = DecimalField<AggregateFunctionStateData>; };
-template <> struct Field::EnumToType<Field::Types::UInt256> { using Type = UInt256; };
-template <> struct Field::EnumToType<Field::Types::Int256>  { using Type = Int256; };
 
-inline constexpr bool isInt64FieldType(Field::Types::Which t)
+inline constexpr bool isInt64OrUInt64FieldType(Field::Types::Which t)
 {
     return t == Field::Types::Int64
         || t == Field::Types::UInt64;
@@ -782,7 +782,7 @@ NearestFieldType<std::decay_t<T>> & Field::get()
     // Disregard signedness when converting between int64 types.
     constexpr Field::Types::Which target = TypeToEnum<StoredType>::value;
     if (target != which
-           && (!isInt64FieldType(target) || !isInt64FieldType(which)))
+           && (!isInt64OrUInt64FieldType(target) || !isInt64OrUInt64FieldType(which)))
         throw Exception(ErrorCodes::LOGICAL_ERROR, "Invalid Field get from type {} to type {}", Types::toString(which), Types::toString(target));
 #endif
 
@@ -834,11 +834,11 @@ T safeGet(Field & field)
     return field.template safeGet<T>();
 }
 
+template <> inline constexpr const char * TypeName<Array> = "Array";
+template <> inline constexpr const char * TypeName<Tuple> = "Tuple";
+template <> inline constexpr const char * TypeName<Map> = "Map";
+template <> inline constexpr const char * TypeName<AggregateFunctionStateData> = "AggregateFunctionState";
 
-template <> struct TypeName<Array> { static std::string get() { return "Array"; } };
-template <> struct TypeName<Tuple> { static std::string get() { return "Tuple"; } };
-template <> struct TypeName<Map>   { static std::string get() { return "Map"; } };
-template <> struct TypeName<AggregateFunctionStateData> { static std::string get() { return "AggregateFunctionState"; } };
 
 template <typename T>
 decltype(auto) castToNearestFieldType(T && x)
@@ -965,7 +965,7 @@ struct fmt::formatter<DB::Field>
 
         /// Only support {}.
         if (it != end && *it != '}')
-            throw format_error("invalid format");
+            throw format_error("Invalid format");
 
         return it;
     }

src/Core/MySQL/MySQLGtid.cpp

@@ -1,6 +1,7 @@
 #include "MySQLGtid.h"
-
 #include <boost/algorithm/string.hpp>
+#include <IO/ReadHelpers.h>
+
 
 namespace DB
 {
@@ -38,7 +39,7 @@ void GTIDSets::parse(const String gtid_format)
         boost::split(server_ids, gset, [](char c) { return c == ':'; });
 
         GTIDSet set;
-        set.uuid = stringToUUID(server_ids[0]);
+        set.uuid = DB::parse<UUID>(server_ids[0]);
 
         for (size_t k = 1; k < server_ids.size(); k++)
         {
@@ -174,8 +175,8 @@ String GTIDSets::toPayload() const
     for (const auto & set : sets)
     {
         // MySQL UUID is big-endian.
-        writeBinaryBigEndian(set.uuid.toUnderType().low, buffer);
-        writeBinaryBigEndian(set.uuid.toUnderType().high, buffer);
+        writeBinaryBigEndian(set.uuid.toUnderType().items[0], buffer);
+        writeBinaryBigEndian(set.uuid.toUnderType().items[1], buffer);
 
         UInt64 intervals_size = set.intervals.size();
         buffer.write(reinterpret_cast<const char *>(&intervals_size), 8);

src/Core/MySQL/MySQLReplication.cpp

@@ -658,12 +658,13 @@ namespace MySQLReplication
         payload.readStrict(reinterpret_cast<char *>(&commit_flag), 1);
 
         // MySQL UUID is big-endian.
-        UInt64 high = 0UL, low = 0UL;
+        UInt64 high = 0UL;
+        UInt64 low = 0UL;
         readBigEndianStrict(payload, reinterpret_cast<char *>(&low), 8);
-        gtid.uuid.toUnderType().low = low;
+        gtid.uuid.toUnderType().items[0] = low;
 
         readBigEndianStrict(payload, reinterpret_cast<char *>(&high), 8);
-        gtid.uuid.toUnderType().high = high;
+        gtid.uuid.toUnderType().items[1] = high;
 
         payload.readStrict(reinterpret_cast<char *>(&gtid.seq_no), 8);
 

src/Core/TypeListNumber.h

@@ -1,20 +1,18 @@
 #pragma once
 
 #include <Core/Types.h>
-#include <Common/UInt128.h>
 #include <Common/TypeList.h>
 
 namespace DB
 {
 
 using TypeListNativeNumbers = TypeList<UInt8, UInt16, UInt32, UInt64, Int8, Int16, Int32, Int64, Float32, Float64>;
-using TypeListExtendedNumbers = TypeList<Int128, UInt256, Int256>;
+using TypeListExtendedNumbers = TypeList<UInt128, Int128, UInt256, Int256>;
 using TypeListDecimalNumbers = TypeList<Decimal32, Decimal64, Decimal128, Decimal256>;
 
 using TypeListGeneralNumbers = typename TypeListConcat<TypeListNativeNumbers, TypeListExtendedNumbers>::Type;
 using TypeListNumbers = typename TypeListConcat<TypeListGeneralNumbers, TypeListDecimalNumbers>::Type;
 
-/// Currently separate because UInt128 cannot be used in every context where other numbers can be used.
-using TypeListNumbersAndUInt128 = typename AppendToTypeList<UInt128, TypeListNumbers>::Type;
+using TypeListNumbersAndUUID = typename TypeListConcat<TypeListNumbers, TypeList<UUID>>::Type;
 
 }

src/Core/Types.h

@@ -3,6 +3,7 @@
 #include <cstdint>
 #include <string>
 #include <vector>
+#include <common/strong_typedef.h>
 #include <common/extended_types.h>
 #include <common/defines.h>
 
@@ -63,60 +64,52 @@ enum class TypeIndex
 #pragma GCC diagnostic pop
 #endif
 
-/// Other int defines are in common/types.h
-using UInt256 = ::wUInt256;
+
+using UInt128 = ::UInt128;
+using UInt256 = ::UInt256;
 using Int128 = ::Int128;
-using Int256 = ::wInt256;
+using Int256 = ::Int256;
 
-/** Note that for types not used in DB, IsNumber is false.
-  */
-template <typename T> constexpr bool IsNumber = false;
+STRONG_TYPEDEF(UInt128, UUID)
 
-template <> inline constexpr bool IsNumber<UInt8> = true;
-template <> inline constexpr bool IsNumber<UInt16> = true;
-template <> inline constexpr bool IsNumber<UInt32> = true;
-template <> inline constexpr bool IsNumber<UInt64> = true;
-template <> inline constexpr bool IsNumber<UInt256> = true;
-template <> inline constexpr bool IsNumber<Int8> = true;
-template <> inline constexpr bool IsNumber<Int16> = true;
-template <> inline constexpr bool IsNumber<Int32> = true;
-template <> inline constexpr bool IsNumber<Int64> = true;
-template <> inline constexpr bool IsNumber<Int128> = true;
-template <> inline constexpr bool IsNumber<Int256> = true;
-template <> inline constexpr bool IsNumber<Float32> = true;
-template <> inline constexpr bool IsNumber<Float64> = true;
 
-template <typename T> struct TypeName;
+template <typename T> constexpr const char * TypeName = "";
 
-template <> struct TypeName<UInt8>   { static constexpr const char * get() { return "UInt8";   } };
-template <> struct TypeName<UInt16>  { static constexpr const char * get() { return "UInt16";  } };
-template <> struct TypeName<UInt32>  { static constexpr const char * get() { return "UInt32";  } };
-template <> struct TypeName<UInt64>  { static constexpr const char * get() { return "UInt64";  } };
-template <> struct TypeName<UInt256> { static constexpr const char * get() { return "UInt256"; } };
-template <> struct TypeName<Int8>    { static constexpr const char * get() { return "Int8";    } };
-template <> struct TypeName<Int16>   { static constexpr const char * get() { return "Int16";   } };
-template <> struct TypeName<Int32>   { static constexpr const char * get() { return "Int32";   } };
-template <> struct TypeName<Int64>   { static constexpr const char * get() { return "Int64";   } };
-template <> struct TypeName<Int128>  { static constexpr const char * get() { return "Int128";  } };
-template <> struct TypeName<Int256> { static constexpr const char * get() { return "Int256";  } };
-template <> struct TypeName<Float32> { static constexpr const char * get() { return "Float32"; } };
-template <> struct TypeName<Float64> { static constexpr const char * get() { return "Float64"; } };
-template <> struct TypeName<String>  { static constexpr const char * get() { return "String";  } };
+template <> inline constexpr const char * TypeName<UInt8> = "UInt8";
+template <> inline constexpr const char * TypeName<UInt16> = "UInt16";
+template <> inline constexpr const char * TypeName<UInt32> = "UInt32";
+template <> inline constexpr const char * TypeName<UInt64> = "UInt64";
+template <> inline constexpr const char * TypeName<UInt128> = "UInt128";
+template <> inline constexpr const char * TypeName<UInt256> = "UInt256";
+template <> inline constexpr const char * TypeName<Int8> = "Int8";
+template <> inline constexpr const char * TypeName<Int16> = "Int16";
+template <> inline constexpr const char * TypeName<Int32> = "Int32";
+template <> inline constexpr const char * TypeName<Int64> = "Int64";
+template <> inline constexpr const char * TypeName<Int128> = "Int128";
+template <> inline constexpr const char * TypeName<Int256> = "Int256";
+template <> inline constexpr const char * TypeName<Float32> = "Float32";
+template <> inline constexpr const char * TypeName<Float64> = "Float64";
+template <> inline constexpr const char * TypeName<String> = "String";
+template <> inline constexpr const char * TypeName<UUID> = "UUID";
+
+/// TODO Try to remove it.
+template <typename T> constexpr TypeIndex TypeId = TypeIndex::Nothing;
+template <> inline constexpr TypeIndex TypeId<UInt8> = TypeIndex::UInt8;
+template <> inline constexpr TypeIndex TypeId<UInt16> = TypeIndex::UInt16;
+template <> inline constexpr TypeIndex TypeId<UInt32> = TypeIndex::UInt32;
+template <> inline constexpr TypeIndex TypeId<UInt64> = TypeIndex::UInt64;
+template <> inline constexpr TypeIndex TypeId<UInt128> = TypeIndex::UInt128;
+template <> inline constexpr TypeIndex TypeId<UInt256> = TypeIndex::UInt256;
+template <> inline constexpr TypeIndex TypeId<Int8> = TypeIndex::Int8;
+template <> inline constexpr TypeIndex TypeId<Int16> = TypeIndex::Int16;
+template <> inline constexpr TypeIndex TypeId<Int32> = TypeIndex::Int32;
+template <> inline constexpr TypeIndex TypeId<Int64> = TypeIndex::Int64;
+template <> inline constexpr TypeIndex TypeId<Int128> = TypeIndex::Int128;
+template <> inline constexpr TypeIndex TypeId<Int256> = TypeIndex::Int256;
+template <> inline constexpr TypeIndex TypeId<Float32> = TypeIndex::Float32;
+template <> inline constexpr TypeIndex TypeId<Float64> = TypeIndex::Float64;
+template <> inline constexpr TypeIndex TypeId<UUID> = TypeIndex::UUID;
 
-template <typename T> struct TypeId;
-template <> struct TypeId<UInt8>    { static constexpr const TypeIndex value = TypeIndex::UInt8;  };
-template <> struct TypeId<UInt16>   { static constexpr const TypeIndex value = TypeIndex::UInt16;  };
-template <> struct TypeId<UInt32>   { static constexpr const TypeIndex value = TypeIndex::UInt32;  };
-template <> struct TypeId<UInt64>   { static constexpr const TypeIndex value = TypeIndex::UInt64;  };
-template <> struct TypeId<UInt256>  { static constexpr const TypeIndex value = TypeIndex::UInt256; };
-template <> struct TypeId<Int8>     { static constexpr const TypeIndex value = TypeIndex::Int8;  };
-template <> struct TypeId<Int16>    { static constexpr const TypeIndex value = TypeIndex::Int16; };
-template <> struct TypeId<Int32>    { static constexpr const TypeIndex value = TypeIndex::Int32; };
-template <> struct TypeId<Int64>    { static constexpr const TypeIndex value = TypeIndex::Int64; };
-template <> struct TypeId<Int128>   { static constexpr const TypeIndex value = TypeIndex::Int128; };
-template <> struct TypeId<Int256>   { static constexpr const TypeIndex value = TypeIndex::Int256; };
-template <> struct TypeId<Float32>  { static constexpr const TypeIndex value = TypeIndex::Float32;  };
-template <> struct TypeId<Float64>  { static constexpr const TypeIndex value = TypeIndex::Float64;  };
 
 /// Not a data type in database, defined just for convenience.
 using Strings = std::vector<String>;
@@ -167,6 +160,12 @@ struct Decimal
     const Decimal<T> & operator /= (const T & x) { value /= x; return *this; }
     const Decimal<T> & operator %= (const T & x) { value %= x; return *this; }
 
+    template <typename U> const Decimal<T> & operator += (const Decimal<U> & x) { value += x.value; return *this; }
+    template <typename U> const Decimal<T> & operator -= (const Decimal<U> & x) { value -= x.value; return *this; }
+    template <typename U> const Decimal<T> & operator *= (const Decimal<U> & x) { value *= x.value; return *this; }
+    template <typename U> const Decimal<T> & operator /= (const Decimal<U> & x) { value /= x.value; return *this; }
+    template <typename U> const Decimal<T> & operator %= (const Decimal<U> & x) { value %= x.value; return *this; }
+
     /// This is to avoid UB for sumWithOverflow()
     void NO_SANITIZE_UNDEFINED addOverflow(const T & x) { value += x; }
 
@@ -175,6 +174,8 @@ struct Decimal
 
 template <typename T> inline bool operator< (const Decimal<T> & x, const Decimal<T> & y) { return x.value < y.value; }
 template <typename T> inline bool operator> (const Decimal<T> & x, const Decimal<T> & y) { return x.value > y.value; }
+template <typename T> inline bool operator<= (const Decimal<T> & x, const Decimal<T> & y) { return x.value <= y.value; }
+template <typename T> inline bool operator>= (const Decimal<T> & x, const Decimal<T> & y) { return x.value >= y.value; }
 template <typename T> inline bool operator== (const Decimal<T> & x, const Decimal<T> & y) { return x.value == y.value; }
 template <typename T> inline bool operator!= (const Decimal<T> & x, const Decimal<T> & y) { return x.value != y.value; }
 
@@ -202,17 +203,17 @@ public:
     {}
 };
 
-template <> struct TypeName<Decimal32>   { static constexpr const char * get() { return "Decimal32";   } };
-template <> struct TypeName<Decimal64>   { static constexpr const char * get() { return "Decimal64";   } };
-template <> struct TypeName<Decimal128>  { static constexpr const char * get() { return "Decimal128";  } };
-template <> struct TypeName<Decimal256>  { static constexpr const char * get() { return "Decimal256";  } };
-template <> struct TypeName<DateTime64>  { static constexpr const char * get() { return "DateTime64";  } };
+template <> inline constexpr const char * TypeName<Decimal32> = "Decimal32";
+template <> inline constexpr const char * TypeName<Decimal64> = "Decimal64";
+template <> inline constexpr const char * TypeName<Decimal128> = "Decimal128";
+template <> inline constexpr const char * TypeName<Decimal256> = "Decimal256";
+template <> inline constexpr const char * TypeName<DateTime64> = "DateTime64";
 
-template <> struct TypeId<Decimal32>    { static constexpr const TypeIndex value = TypeIndex::Decimal32; };
-template <> struct TypeId<Decimal64>    { static constexpr const TypeIndex value = TypeIndex::Decimal64; };
-template <> struct TypeId<Decimal128>   { static constexpr const TypeIndex value = TypeIndex::Decimal128; };
-template <> struct TypeId<Decimal256>   { static constexpr const TypeIndex value = TypeIndex::Decimal256; };
-template <> struct TypeId<DateTime64>   { static constexpr const TypeIndex value = TypeIndex::DateTime64; };
+template <> inline constexpr TypeIndex TypeId<Decimal32> = TypeIndex::Decimal32;
+template <> inline constexpr TypeIndex TypeId<Decimal64> = TypeIndex::Decimal64;
+template <> inline constexpr TypeIndex TypeId<Decimal128> = TypeIndex::Decimal128;
+template <> inline constexpr TypeIndex TypeId<Decimal256> = TypeIndex::Decimal256;
+template <> inline constexpr TypeIndex TypeId<DateTime64> = TypeIndex::DateTime64;
 
 template <typename T> constexpr bool IsDecimalNumber = false;
 template <> inline constexpr bool IsDecimalNumber<Decimal32> = true;
@@ -229,8 +230,11 @@ template <> struct NativeType<Decimal256> { using Type = Int256; };
 template <> struct NativeType<DateTime64> { using Type = Int64; };
 
 template <typename T> constexpr bool OverBigInt = false;
+template <> inline constexpr bool OverBigInt<Int128> = true;
+template <> inline constexpr bool OverBigInt<UInt128> = true;
 template <> inline constexpr bool OverBigInt<Int256> = true;
 template <> inline constexpr bool OverBigInt<UInt256> = true;
+template <> inline constexpr bool OverBigInt<Decimal128> = true;
 template <> inline constexpr bool OverBigInt<Decimal256> = true;
 
 inline constexpr const char * getTypeName(TypeIndex idx)
@@ -238,31 +242,31 @@ inline constexpr const char * getTypeName(TypeIndex idx)
     switch (idx)
     {
         case TypeIndex::Nothing:    return "Nothing";
-        case TypeIndex::UInt8:      return TypeName<UInt8>::get();
-        case TypeIndex::UInt16:     return TypeName<UInt16>::get();
-        case TypeIndex::UInt32:     return TypeName<UInt32>::get();
-        case TypeIndex::UInt64:     return TypeName<UInt64>::get();
+        case TypeIndex::UInt8:      return "UInt8";
+        case TypeIndex::UInt16:     return "UInt16";
+        case TypeIndex::UInt32:     return "UInt32";
+        case TypeIndex::UInt64:     return "UInt64";
         case TypeIndex::UInt128:    return "UInt128";
-        case TypeIndex::UInt256:    return TypeName<UInt256>::get();
-        case TypeIndex::Int8:       return TypeName<Int8>::get();
-        case TypeIndex::Int16:      return TypeName<Int16>::get();
-        case TypeIndex::Int32:      return TypeName<Int32>::get();
-        case TypeIndex::Int64:      return TypeName<Int64>::get();
-        case TypeIndex::Int128:     return TypeName<Int128>::get();
-        case TypeIndex::Int256:     return TypeName<Int256>::get();
-        case TypeIndex::Float32:    return TypeName<Float32>::get();
-        case TypeIndex::Float64:    return TypeName<Float64>::get();
+        case TypeIndex::UInt256:    return "UInt256";
+        case TypeIndex::Int8:       return "Int8";
+        case TypeIndex::Int16:      return "Int16";
+        case TypeIndex::Int32:      return "Int32";
+        case TypeIndex::Int64:      return "Int64";
+        case TypeIndex::Int128:     return "Int128";
+        case TypeIndex::Int256:     return "Int256";
+        case TypeIndex::Float32:    return "Float32";
+        case TypeIndex::Float64:    return "Float64";
         case TypeIndex::Date:       return "Date";
         case TypeIndex::DateTime:   return "DateTime";
         case TypeIndex::DateTime64: return "DateTime64";
-        case TypeIndex::String:     return TypeName<String>::get();
+        case TypeIndex::String:     return "String";
         case TypeIndex::FixedString: return "FixedString";
         case TypeIndex::Enum8:      return "Enum8";
         case TypeIndex::Enum16:     return "Enum16";
-        case TypeIndex::Decimal32:  return TypeName<Decimal32>::get();
-        case TypeIndex::Decimal64:  return TypeName<Decimal64>::get();
-        case TypeIndex::Decimal128: return TypeName<Decimal128>::get();
-        case TypeIndex::Decimal256: return TypeName<Decimal256>::get();
+        case TypeIndex::Decimal32:  return "Decimal32";
+        case TypeIndex::Decimal64:  return "Decimal64";
+        case TypeIndex::Decimal128: return "Decimal128";
+        case TypeIndex::Decimal256: return "Decimal256";
         case TypeIndex::UUID:       return "UUID";
         case TypeIndex::Array:      return "Array";
         case TypeIndex::Tuple:      return "Tuple";

src/Core/UUID.cpp

@@ -0,0 +1,19 @@
+#include <Core/UUID.h>
+#include <Common/thread_local_rng.h>
+
+
+namespace DB
+{
+
+namespace UUIDHelpers
+{
+    UUID generateV4()
+    {
+        UInt128 res{thread_local_rng(), thread_local_rng()};
+        res.items[0] = (res.items[0] & 0xffffffffffff0fffull) | 0x0000000000004000ull;
+        res.items[1] = (res.items[1] & 0x3fffffffffffffffull) | 0x8000000000000000ull;
+        return UUID{res};
+    }
+}
+
+}

src/Core/UUID.h

@@ -1,25 +1,17 @@
 #pragma once
 
-#include <Common/UInt128.h>
-#include <common/strong_typedef.h>
-#include <Common/thread_local_rng.h>
+#include <Core/Types.h>
+
 
 namespace DB
 {
 
-STRONG_TYPEDEF(UInt128, UUID)
-
 namespace UUIDHelpers
 {
-    inline UUID generateV4()
-    {
-        UInt128 res{thread_local_rng(), thread_local_rng()};
-        res.low = (res.low & 0xffffffffffff0fffull) | 0x0000000000004000ull;
-        res.high = (res.high & 0x3fffffffffffffffull) | 0x8000000000000000ull;
-        return UUID{res};
-    }
+    /// Generate random UUID.
+    UUID generateV4();
 
-    const UUID Nil = UUID(UInt128(0, 0));
+    const UUID Nil{};
 }
 
 }