mirror of
https://github.com/ClickHouse/ClickHouse.git
synced 2024-12-15 10:52:30 +00:00
1355 lines
56 KiB
C++
1355 lines
56 KiB
C++
#pragma once
|
|
|
|
#include <Common/memcmpSmall.h>
|
|
#include <Common/assert_cast.h>
|
|
#include <Common/TargetSpecific.h>
|
|
|
|
#include <Columns/ColumnsNumber.h>
|
|
#include <Columns/ColumnConst.h>
|
|
#include <Columns/ColumnDecimal.h>
|
|
#include <Columns/ColumnString.h>
|
|
#include <Columns/ColumnFixedString.h>
|
|
#include <Columns/ColumnTuple.h>
|
|
#include <Columns/ColumnArray.h>
|
|
|
|
#include <DataTypes/DataTypeDate.h>
|
|
#include <DataTypes/DataTypeDateTime.h>
|
|
#include <DataTypes/DataTypeDateTime64.h>
|
|
#include <DataTypes/DataTypeEnum.h>
|
|
#include <DataTypes/DataTypeFixedString.h>
|
|
#include <DataTypes/DataTypeNothing.h>
|
|
#include <DataTypes/DataTypeNullable.h>
|
|
#include <DataTypes/DataTypeString.h>
|
|
#include <DataTypes/DataTypeTuple.h>
|
|
#include <DataTypes/DataTypeUUID.h>
|
|
#include <DataTypes/DataTypesNumber.h>
|
|
#include <DataTypes/getLeastSupertype.h>
|
|
|
|
#include <Interpreters/convertFieldToType.h>
|
|
#include <Interpreters/castColumn.h>
|
|
|
|
#include <Functions/IFunctionAdaptors.h>
|
|
#include <Functions/FunctionHelpers.h>
|
|
#include <Functions/IsOperation.h>
|
|
|
|
#include <Core/AccurateComparison.h>
|
|
#include <Core/DecimalComparison.h>
|
|
|
|
#include <IO/ReadBufferFromMemory.h>
|
|
#include <IO/ReadHelpers.h>
|
|
|
|
#include <limits>
|
|
#include <type_traits>
|
|
|
|
#if USE_EMBEDDED_COMPILER
|
|
#include <DataTypes/Native.h>
|
|
|
|
#pragma GCC diagnostic push
|
|
#pragma GCC diagnostic ignored "-Wunused-parameter"
|
|
#include <llvm/IR/IRBuilder.h>
|
|
#pragma GCC diagnostic pop
|
|
#endif
|
|
|
|
|
|
namespace DB
|
|
{
|
|
|
|
namespace ErrorCodes
|
|
{
|
|
extern const int ILLEGAL_COLUMN;
|
|
extern const int ILLEGAL_TYPE_OF_ARGUMENT;
|
|
extern const int LOGICAL_ERROR;
|
|
extern const int NOT_IMPLEMENTED;
|
|
extern const int BAD_ARGUMENTS;
|
|
}
|
|
|
|
|
|
/** Comparison functions: ==, !=, <, >, <=, >=.
|
|
* The comparison functions always return 0 or 1 (UInt8).
|
|
*
|
|
* You can compare the following types:
|
|
* - numbers and decimals;
|
|
* - strings and fixed strings;
|
|
* - dates;
|
|
* - datetimes;
|
|
* within each group, but not from different groups;
|
|
* - tuples (lexicographic comparison).
|
|
*
|
|
* Exception: You can compare the date and datetime with a constant string. Example: EventDate = '2015-01-01'.
|
|
*/
|
|
|
|
|
|
template <typename A, typename B, typename Op>
|
|
struct NumComparisonImpl
|
|
{
|
|
using ContainerA = PaddedPODArray<A>;
|
|
using ContainerB = PaddedPODArray<B>;
|
|
|
|
MULTITARGET_FUNCTION_AVX2_SSE42(
|
|
MULTITARGET_FUNCTION_HEADER(static void), vectorVectorImpl, MULTITARGET_FUNCTION_BODY(( /// NOLINT
|
|
const ContainerA & a, const ContainerB & b, PaddedPODArray<UInt8> & c)
|
|
{
|
|
/** GCC 4.8.2 vectorizes a loop only if it is written in this form.
|
|
* In this case, if you loop through the array index (the code will look simpler),
|
|
* the loop will not be vectorized.
|
|
*/
|
|
|
|
size_t size = a.size();
|
|
const A * __restrict a_pos = a.data();
|
|
const B * __restrict b_pos = b.data();
|
|
UInt8 * __restrict c_pos = c.data();
|
|
const A * a_end = a_pos + size;
|
|
|
|
while (a_pos < a_end)
|
|
{
|
|
*c_pos = Op::apply(*a_pos, *b_pos);
|
|
++a_pos;
|
|
++b_pos;
|
|
++c_pos;
|
|
}
|
|
}))
|
|
|
|
static void NO_INLINE vectorVector(const ContainerA & a, const ContainerB & b, PaddedPODArray<UInt8> & c)
|
|
{
|
|
#if USE_MULTITARGET_CODE
|
|
if (isArchSupported(TargetArch::AVX2))
|
|
{
|
|
vectorVectorImplAVX2(a, b, c);
|
|
return;
|
|
}
|
|
else if (isArchSupported(TargetArch::SSE42))
|
|
{
|
|
vectorVectorImplSSE42(a, b, c);
|
|
return;
|
|
}
|
|
#endif
|
|
|
|
vectorVectorImpl(a, b, c);
|
|
}
|
|
|
|
|
|
MULTITARGET_FUNCTION_AVX2_SSE42(
|
|
MULTITARGET_FUNCTION_HEADER(static void), vectorConstantImpl, MULTITARGET_FUNCTION_BODY(( /// NOLINT
|
|
const ContainerA & a, B b, PaddedPODArray<UInt8> & c)
|
|
{
|
|
size_t size = a.size();
|
|
const A * __restrict a_pos = a.data();
|
|
UInt8 * __restrict c_pos = c.data();
|
|
const A * a_end = a_pos + size;
|
|
|
|
while (a_pos < a_end)
|
|
{
|
|
*c_pos = Op::apply(*a_pos, b);
|
|
++a_pos;
|
|
++c_pos;
|
|
}
|
|
}))
|
|
|
|
static void NO_INLINE vectorConstant(const ContainerA & a, B b, PaddedPODArray<UInt8> & c)
|
|
{
|
|
#if USE_MULTITARGET_CODE
|
|
if (isArchSupported(TargetArch::AVX2))
|
|
{
|
|
vectorConstantImplAVX2(a, b, c);
|
|
return;
|
|
}
|
|
else if (isArchSupported(TargetArch::SSE42))
|
|
{
|
|
vectorConstantImplSSE42(a, b, c);
|
|
return;
|
|
}
|
|
#endif
|
|
|
|
vectorConstantImpl(a, b, c);
|
|
}
|
|
|
|
static void constantVector(A a, const ContainerB & b, PaddedPODArray<UInt8> & c)
|
|
{
|
|
NumComparisonImpl<B, A, typename Op::SymmetricOp>::vectorConstant(b, a, c);
|
|
}
|
|
|
|
static void constantConstant(A a, B b, UInt8 & c)
|
|
{
|
|
c = Op::apply(a, b);
|
|
}
|
|
};
|
|
|
|
|
|
template <typename Op>
|
|
struct StringComparisonImpl
|
|
{
|
|
static void NO_INLINE string_vector_string_vector( /// NOLINT
|
|
const ColumnString::Chars & a_data, const ColumnString::Offsets & a_offsets,
|
|
const ColumnString::Chars & b_data, const ColumnString::Offsets & b_offsets,
|
|
PaddedPODArray<UInt8> & c)
|
|
{
|
|
size_t size = a_offsets.size();
|
|
ColumnString::Offset prev_a_offset = 0;
|
|
ColumnString::Offset prev_b_offset = 0;
|
|
|
|
for (size_t i = 0; i < size; ++i)
|
|
{
|
|
c[i] = Op::apply(memcmpSmallAllowOverflow15(
|
|
a_data.data() + prev_a_offset, a_offsets[i] - prev_a_offset - 1,
|
|
b_data.data() + prev_b_offset, b_offsets[i] - prev_b_offset - 1), 0);
|
|
|
|
prev_a_offset = a_offsets[i];
|
|
prev_b_offset = b_offsets[i];
|
|
}
|
|
}
|
|
|
|
static void NO_INLINE string_vector_fixed_string_vector( /// NOLINT
|
|
const ColumnString::Chars & a_data, const ColumnString::Offsets & a_offsets,
|
|
const ColumnString::Chars & b_data, ColumnString::Offset b_n,
|
|
PaddedPODArray<UInt8> & c)
|
|
{
|
|
size_t size = a_offsets.size();
|
|
ColumnString::Offset prev_a_offset = 0;
|
|
|
|
for (size_t i = 0; i < size; ++i)
|
|
{
|
|
c[i] = Op::apply(memcmpSmallLikeZeroPaddedAllowOverflow15(
|
|
a_data.data() + prev_a_offset, a_offsets[i] - prev_a_offset - 1,
|
|
b_data.data() + i * b_n, b_n), 0);
|
|
|
|
prev_a_offset = a_offsets[i];
|
|
}
|
|
}
|
|
|
|
static void NO_INLINE string_vector_constant( /// NOLINT
|
|
const ColumnString::Chars & a_data, const ColumnString::Offsets & a_offsets,
|
|
const ColumnString::Chars & b_data, ColumnString::Offset b_size,
|
|
PaddedPODArray<UInt8> & c)
|
|
{
|
|
size_t size = a_offsets.size();
|
|
ColumnString::Offset prev_a_offset = 0;
|
|
|
|
for (size_t i = 0; i < size; ++i)
|
|
{
|
|
c[i] = Op::apply(memcmpSmallAllowOverflow15(
|
|
a_data.data() + prev_a_offset, a_offsets[i] - prev_a_offset - 1,
|
|
b_data.data(), b_size), 0);
|
|
|
|
prev_a_offset = a_offsets[i];
|
|
}
|
|
}
|
|
|
|
static void fixed_string_vector_string_vector( /// NOLINT
|
|
const ColumnString::Chars & a_data, ColumnString::Offset a_n,
|
|
const ColumnString::Chars & b_data, const ColumnString::Offsets & b_offsets,
|
|
PaddedPODArray<UInt8> & c)
|
|
{
|
|
StringComparisonImpl<typename Op::SymmetricOp>::string_vector_fixed_string_vector(b_data, b_offsets, a_data, a_n, c);
|
|
}
|
|
|
|
static void NO_INLINE fixed_string_vector_fixed_string_vector_16( /// NOLINT
|
|
const ColumnString::Chars & a_data,
|
|
const ColumnString::Chars & b_data,
|
|
PaddedPODArray<UInt8> & c)
|
|
{
|
|
size_t size = a_data.size();
|
|
|
|
for (size_t i = 0, j = 0; i < size; i += 16, ++j)
|
|
c[j] = Op::apply(memcmp16(&a_data[i], &b_data[i]), 0);
|
|
}
|
|
|
|
static void NO_INLINE fixed_string_vector_constant_16( /// NOLINT
|
|
const ColumnString::Chars & a_data,
|
|
const ColumnString::Chars & b_data,
|
|
PaddedPODArray<UInt8> & c)
|
|
{
|
|
size_t size = a_data.size();
|
|
|
|
for (size_t i = 0, j = 0; i < size; i += 16, ++j)
|
|
c[j] = Op::apply(memcmp16(&a_data[i], &b_data[0]), 0);
|
|
}
|
|
|
|
static void NO_INLINE fixed_string_vector_fixed_string_vector( /// NOLINT
|
|
const ColumnString::Chars & a_data, ColumnString::Offset a_n,
|
|
const ColumnString::Chars & b_data, ColumnString::Offset b_n,
|
|
PaddedPODArray<UInt8> & c)
|
|
{
|
|
if (a_n == 16 && b_n == 16)
|
|
{
|
|
/** Specialization if both sizes are 16.
|
|
* To more efficient comparison of IPv6 addresses stored in FixedString(16).
|
|
*/
|
|
fixed_string_vector_fixed_string_vector_16(a_data, b_data, c);
|
|
}
|
|
else if (a_n == b_n)
|
|
{
|
|
size_t size = a_data.size();
|
|
for (size_t i = 0, j = 0; i < size; i += a_n, ++j)
|
|
c[j] = Op::apply(memcmpSmallAllowOverflow15(a_data.data() + i, b_data.data() + i, a_n), 0);
|
|
}
|
|
else
|
|
{
|
|
size_t size = a_data.size() / a_n;
|
|
|
|
for (size_t i = 0; i < size; ++i)
|
|
c[i] = Op::apply(memcmpSmallLikeZeroPaddedAllowOverflow15(a_data.data() + i * a_n, a_n, b_data.data() + i * b_n, b_n), 0);
|
|
}
|
|
}
|
|
|
|
static void NO_INLINE fixed_string_vector_constant( /// NOLINT
|
|
const ColumnString::Chars & a_data, ColumnString::Offset a_n,
|
|
const ColumnString::Chars & b_data, ColumnString::Offset b_size,
|
|
PaddedPODArray<UInt8> & c)
|
|
{
|
|
if (a_n == 16 && b_size == 16)
|
|
{
|
|
fixed_string_vector_constant_16(a_data, b_data, c);
|
|
}
|
|
else if (a_n == b_size)
|
|
{
|
|
size_t size = a_data.size();
|
|
for (size_t i = 0, j = 0; i < size; i += a_n, ++j)
|
|
c[j] = Op::apply(memcmpSmallAllowOverflow15(a_data.data() + i, b_data.data(), a_n), 0);
|
|
}
|
|
else
|
|
{
|
|
size_t size = a_data.size();
|
|
for (size_t i = 0, j = 0; i < size; i += a_n, ++j)
|
|
c[j] = Op::apply(memcmpSmallLikeZeroPaddedAllowOverflow15(a_data.data() + i, a_n, b_data.data(), b_size), 0);
|
|
}
|
|
}
|
|
|
|
static void constant_string_vector( /// NOLINT
|
|
const ColumnString::Chars & a_data, ColumnString::Offset a_size,
|
|
const ColumnString::Chars & b_data, const ColumnString::Offsets & b_offsets,
|
|
PaddedPODArray<UInt8> & c)
|
|
{
|
|
StringComparisonImpl<typename Op::SymmetricOp>::string_vector_constant(b_data, b_offsets, a_data, a_size, c);
|
|
}
|
|
|
|
static void constant_fixed_string_vector( /// NOLINT
|
|
const ColumnString::Chars & a_data, ColumnString::Offset a_size,
|
|
const ColumnString::Chars & b_data, ColumnString::Offset b_n,
|
|
PaddedPODArray<UInt8> & c)
|
|
{
|
|
StringComparisonImpl<typename Op::SymmetricOp>::fixed_string_vector_constant(b_data, b_n, a_data, a_size, c);
|
|
}
|
|
};
|
|
|
|
|
|
/// Comparisons for equality/inequality are implemented slightly more efficient.
|
|
template <bool positive>
|
|
struct StringEqualsImpl
|
|
{
|
|
static void NO_INLINE string_vector_string_vector( /// NOLINT
|
|
const ColumnString::Chars & a_data, const ColumnString::Offsets & a_offsets,
|
|
const ColumnString::Chars & b_data, const ColumnString::Offsets & b_offsets,
|
|
PaddedPODArray<UInt8> & c)
|
|
{
|
|
size_t size = a_offsets.size();
|
|
ColumnString::Offset prev_a_offset = 0;
|
|
ColumnString::Offset prev_b_offset = 0;
|
|
|
|
for (size_t i = 0; i < size; ++i)
|
|
{
|
|
auto a_size = a_offsets[i] - prev_a_offset - 1;
|
|
auto b_size = b_offsets[i] - prev_b_offset - 1;
|
|
|
|
c[i] = positive == memequalSmallAllowOverflow15(
|
|
a_data.data() + prev_a_offset, a_size,
|
|
b_data.data() + prev_b_offset, b_size);
|
|
|
|
prev_a_offset = a_offsets[i];
|
|
prev_b_offset = b_offsets[i];
|
|
}
|
|
}
|
|
|
|
static void NO_INLINE string_vector_fixed_string_vector( /// NOLINT
|
|
const ColumnString::Chars & a_data, const ColumnString::Offsets & a_offsets,
|
|
const ColumnString::Chars & b_data, ColumnString::Offset b_n,
|
|
PaddedPODArray<UInt8> & c)
|
|
{
|
|
size_t size = a_offsets.size();
|
|
ColumnString::Offset prev_a_offset = 0;
|
|
|
|
for (size_t i = 0; i < size; ++i)
|
|
{
|
|
auto a_size = a_offsets[i] - prev_a_offset - 1;
|
|
|
|
c[i] = positive == memequalSmallLikeZeroPaddedAllowOverflow15(
|
|
a_data.data() + prev_a_offset, a_size,
|
|
b_data.data() + b_n * i, b_n);
|
|
|
|
prev_a_offset = a_offsets[i];
|
|
}
|
|
}
|
|
|
|
static void NO_INLINE string_vector_constant( /// NOLINT
|
|
const ColumnString::Chars & a_data, const ColumnString::Offsets & a_offsets,
|
|
const ColumnString::Chars & b_data, ColumnString::Offset b_size,
|
|
PaddedPODArray<UInt8> & c)
|
|
{
|
|
size_t size = a_offsets.size();
|
|
ColumnString::Offset prev_a_offset = 0;
|
|
|
|
for (size_t i = 0; i < size; ++i)
|
|
{
|
|
auto a_size = a_offsets[i] - prev_a_offset - 1;
|
|
|
|
c[i] = positive == memequalSmallAllowOverflow15(
|
|
a_data.data() + prev_a_offset, a_size,
|
|
b_data.data(), b_size);
|
|
|
|
prev_a_offset = a_offsets[i];
|
|
}
|
|
}
|
|
|
|
static void NO_INLINE fixed_string_vector_fixed_string_vector_16( /// NOLINT
|
|
const ColumnString::Chars & a_data,
|
|
const ColumnString::Chars & b_data,
|
|
PaddedPODArray<UInt8> & c)
|
|
{
|
|
size_t size = a_data.size() / 16;
|
|
|
|
for (size_t i = 0; i < size; ++i)
|
|
c[i] = positive == memequal16(
|
|
a_data.data() + i * 16,
|
|
b_data.data() + i * 16);
|
|
}
|
|
|
|
static void NO_INLINE fixed_string_vector_constant_16( /// NOLINT
|
|
const ColumnString::Chars & a_data,
|
|
const ColumnString::Chars & b_data,
|
|
PaddedPODArray<UInt8> & c)
|
|
{
|
|
size_t size = a_data.size() / 16;
|
|
|
|
for (size_t i = 0; i < size; ++i)
|
|
c[i] = positive == memequal16(
|
|
a_data.data() + i * 16,
|
|
b_data.data());
|
|
}
|
|
|
|
static void NO_INLINE fixed_string_vector_fixed_string_vector( /// NOLINT
|
|
const ColumnString::Chars & a_data, ColumnString::Offset a_n,
|
|
const ColumnString::Chars & b_data, ColumnString::Offset b_n,
|
|
PaddedPODArray<UInt8> & c)
|
|
{
|
|
/** Specialization if both sizes are 16.
|
|
* To more efficient comparison of IPv6 addresses stored in FixedString(16).
|
|
*/
|
|
if (a_n == 16 && b_n == 16)
|
|
{
|
|
fixed_string_vector_fixed_string_vector_16(a_data, b_data, c);
|
|
}
|
|
else if (a_n == b_n)
|
|
{
|
|
size_t size = a_data.size() / a_n;
|
|
for (size_t i = 0; i < size; ++i)
|
|
c[i] = positive == memequalSmallAllowOverflow15(a_data.data() + i * a_n, a_n, b_data.data() + i * a_n, a_n);
|
|
}
|
|
else
|
|
{
|
|
size_t size = a_data.size() / a_n;
|
|
for (size_t i = 0; i < size; ++i)
|
|
c[i] = positive == memequalSmallLikeZeroPaddedAllowOverflow15(a_data.data() + i * a_n, a_n, b_data.data() + i * b_n, b_n);
|
|
}
|
|
}
|
|
|
|
static void NO_INLINE fixed_string_vector_constant( /// NOLINT
|
|
const ColumnString::Chars & a_data, ColumnString::Offset a_n,
|
|
const ColumnString::Chars & b_data, ColumnString::Offset b_size,
|
|
PaddedPODArray<UInt8> & c)
|
|
{
|
|
if (a_n == 16 && b_size == 16)
|
|
{
|
|
fixed_string_vector_constant_16(a_data, b_data, c);
|
|
}
|
|
else
|
|
{
|
|
size_t size = a_data.size() / a_n;
|
|
for (size_t i = 0; i < size; ++i)
|
|
c[i] = positive == memequalSmallLikeZeroPaddedAllowOverflow15(a_data.data() + i * a_n, a_n, b_data.data(), b_size);
|
|
}
|
|
}
|
|
|
|
static void fixed_string_vector_string_vector( /// NOLINT
|
|
const ColumnString::Chars & a_data, ColumnString::Offset a_n,
|
|
const ColumnString::Chars & b_data, const ColumnString::Offsets & b_offsets,
|
|
PaddedPODArray<UInt8> & c)
|
|
{
|
|
string_vector_fixed_string_vector(b_data, b_offsets, a_data, a_n, c);
|
|
}
|
|
|
|
static void constant_string_vector( /// NOLINT
|
|
const ColumnString::Chars & a_data, ColumnString::Offset a_size,
|
|
const ColumnString::Chars & b_data, const ColumnString::Offsets & b_offsets,
|
|
PaddedPODArray<UInt8> & c)
|
|
{
|
|
string_vector_constant(b_data, b_offsets, a_data, a_size, c);
|
|
}
|
|
|
|
static void constant_fixed_string_vector( /// NOLINT
|
|
const ColumnString::Chars & a_data, ColumnString::Offset a_size,
|
|
const ColumnString::Chars & b_data, ColumnString::Offset b_n,
|
|
PaddedPODArray<UInt8> & c)
|
|
{
|
|
fixed_string_vector_constant(b_data, b_n, a_data, a_size, c);
|
|
}
|
|
};
|
|
|
|
|
|
template <typename A, typename B>
|
|
struct StringComparisonImpl<EqualsOp<A, B>> : StringEqualsImpl<true> {};
|
|
|
|
template <typename A, typename B>
|
|
struct StringComparisonImpl<NotEqualsOp<A, B>> : StringEqualsImpl<false> {};
|
|
|
|
|
|
/// Generic version, implemented for columns of same type.
|
|
template <typename Op>
|
|
struct GenericComparisonImpl
|
|
{
|
|
static void NO_INLINE vectorVector(const IColumn & a, const IColumn & b, PaddedPODArray<UInt8> & c)
|
|
{
|
|
for (size_t i = 0, size = a.size(); i < size; ++i)
|
|
c[i] = Op::apply(a.compareAt(i, i, b, 1), 0);
|
|
}
|
|
|
|
static void NO_INLINE vectorConstant(const IColumn & a, const IColumn & b, PaddedPODArray<UInt8> & c)
|
|
{
|
|
auto b_materialized = b.cloneResized(1)->convertToFullColumnIfConst();
|
|
for (size_t i = 0, size = a.size(); i < size; ++i)
|
|
c[i] = Op::apply(a.compareAt(i, 0, *b_materialized, 1), 0);
|
|
}
|
|
|
|
static void constantVector(const IColumn & a, const IColumn & b, PaddedPODArray<UInt8> & c)
|
|
{
|
|
GenericComparisonImpl<typename Op::SymmetricOp>::vectorConstant(b, a, c);
|
|
}
|
|
|
|
static void constantConstant(const IColumn & a, const IColumn & b, UInt8 & c)
|
|
{
|
|
c = Op::apply(a.compareAt(0, 0, b, 1), 0);
|
|
}
|
|
};
|
|
|
|
|
|
#if USE_EMBEDDED_COMPILER
|
|
|
|
template <template <typename, typename> typename Op> struct CompileOp;
|
|
|
|
template <> struct CompileOp<EqualsOp>
|
|
{
|
|
static llvm::Value * compile(llvm::IRBuilder<> & b, llvm::Value * x, llvm::Value * y, bool /*is_signed*/)
|
|
{
|
|
return x->getType()->isIntegerTy() ? b.CreateICmpEQ(x, y) : b.CreateFCmpOEQ(x, y); /// qNaNs always compare false
|
|
}
|
|
};
|
|
|
|
template <> struct CompileOp<NotEqualsOp>
|
|
{
|
|
static llvm::Value * compile(llvm::IRBuilder<> & b, llvm::Value * x, llvm::Value * y, bool /*is_signed*/)
|
|
{
|
|
return x->getType()->isIntegerTy() ? b.CreateICmpNE(x, y) : b.CreateFCmpONE(x, y);
|
|
}
|
|
};
|
|
|
|
template <> struct CompileOp<LessOp>
|
|
{
|
|
static llvm::Value * compile(llvm::IRBuilder<> & b, llvm::Value * x, llvm::Value * y, bool is_signed)
|
|
{
|
|
return x->getType()->isIntegerTy() ? (is_signed ? b.CreateICmpSLT(x, y) : b.CreateICmpULT(x, y)) : b.CreateFCmpOLT(x, y);
|
|
}
|
|
};
|
|
|
|
template <> struct CompileOp<GreaterOp>
|
|
{
|
|
static llvm::Value * compile(llvm::IRBuilder<> & b, llvm::Value * x, llvm::Value * y, bool is_signed)
|
|
{
|
|
return x->getType()->isIntegerTy() ? (is_signed ? b.CreateICmpSGT(x, y) : b.CreateICmpUGT(x, y)) : b.CreateFCmpOGT(x, y);
|
|
}
|
|
};
|
|
|
|
template <> struct CompileOp<LessOrEqualsOp>
|
|
{
|
|
static llvm::Value * compile(llvm::IRBuilder<> & b, llvm::Value * x, llvm::Value * y, bool is_signed)
|
|
{
|
|
return x->getType()->isIntegerTy() ? (is_signed ? b.CreateICmpSLE(x, y) : b.CreateICmpULE(x, y)) : b.CreateFCmpOLE(x, y);
|
|
}
|
|
};
|
|
|
|
template <> struct CompileOp<GreaterOrEqualsOp>
|
|
{
|
|
static llvm::Value * compile(llvm::IRBuilder<> & b, llvm::Value * x, llvm::Value * y, bool is_signed)
|
|
{
|
|
return x->getType()->isIntegerTy() ? (is_signed ? b.CreateICmpSGE(x, y) : b.CreateICmpUGE(x, y)) : b.CreateFCmpOGE(x, y);
|
|
}
|
|
};
|
|
|
|
#endif
|
|
|
|
|
|
struct NameEquals { static constexpr auto name = "equals"; };
|
|
struct NameNotEquals { static constexpr auto name = "notEquals"; };
|
|
struct NameLess { static constexpr auto name = "less"; };
|
|
struct NameGreater { static constexpr auto name = "greater"; };
|
|
struct NameLessOrEquals { static constexpr auto name = "lessOrEquals"; };
|
|
struct NameGreaterOrEquals { static constexpr auto name = "greaterOrEquals"; };
|
|
|
|
|
|
template <template <typename, typename> class Op, typename Name>
|
|
class FunctionComparison : public IFunction
|
|
{
|
|
public:
|
|
static constexpr auto name = Name::name;
|
|
static FunctionPtr create(ContextPtr context) { return std::make_shared<FunctionComparison>(context); }
|
|
|
|
explicit FunctionComparison(ContextPtr context_)
|
|
: context(context_), check_decimal_overflow(decimalCheckComparisonOverflow(context)) {}
|
|
|
|
private:
|
|
ContextPtr context;
|
|
bool check_decimal_overflow = true;
|
|
|
|
template <typename T0, typename T1>
|
|
ColumnPtr executeNumRightType(const ColumnVector<T0> * col_left, const IColumn * col_right_untyped) const
|
|
{
|
|
if (const ColumnVector<T1> * col_right = checkAndGetColumn<ColumnVector<T1>>(col_right_untyped))
|
|
{
|
|
auto col_res = ColumnUInt8::create();
|
|
|
|
ColumnUInt8::Container & vec_res = col_res->getData();
|
|
vec_res.resize(col_left->getData().size());
|
|
NumComparisonImpl<T0, T1, Op<T0, T1>>::vectorVector(col_left->getData(), col_right->getData(), vec_res);
|
|
|
|
return col_res;
|
|
}
|
|
else if (auto col_right_const = checkAndGetColumnConst<ColumnVector<T1>>(col_right_untyped))
|
|
{
|
|
auto col_res = ColumnUInt8::create();
|
|
|
|
ColumnUInt8::Container & vec_res = col_res->getData();
|
|
vec_res.resize(col_left->size());
|
|
NumComparisonImpl<T0, T1, Op<T0, T1>>::vectorConstant(col_left->getData(), col_right_const->template getValue<T1>(), vec_res);
|
|
|
|
return col_res;
|
|
}
|
|
|
|
return nullptr;
|
|
}
|
|
|
|
template <typename T0, typename T1>
|
|
ColumnPtr executeNumConstRightType(const ColumnConst * col_left, const IColumn * col_right_untyped) const
|
|
{
|
|
if (const ColumnVector<T1> * col_right = checkAndGetColumn<ColumnVector<T1>>(col_right_untyped))
|
|
{
|
|
auto col_res = ColumnUInt8::create();
|
|
|
|
ColumnUInt8::Container & vec_res = col_res->getData();
|
|
vec_res.resize(col_left->size());
|
|
NumComparisonImpl<T0, T1, Op<T0, T1>>::constantVector(col_left->template getValue<T0>(), col_right->getData(), vec_res);
|
|
|
|
return col_res;
|
|
}
|
|
else if (auto col_right_const = checkAndGetColumnConst<ColumnVector<T1>>(col_right_untyped))
|
|
{
|
|
UInt8 res = 0;
|
|
NumComparisonImpl<T0, T1, Op<T0, T1>>::constantConstant(col_left->template getValue<T0>(), col_right_const->template getValue<T1>(), res);
|
|
|
|
return DataTypeUInt8().createColumnConst(col_left->size(), toField(res));
|
|
}
|
|
|
|
return nullptr;
|
|
}
|
|
|
|
template <typename T0>
|
|
ColumnPtr executeNumLeftType(const IColumn * col_left_untyped, const IColumn * col_right_untyped) const
|
|
{
|
|
ColumnPtr res = nullptr;
|
|
if (const ColumnVector<T0> * col_left = checkAndGetColumn<ColumnVector<T0>>(col_left_untyped))
|
|
{
|
|
if ( (res = executeNumRightType<T0, UInt8>(col_left, col_right_untyped))
|
|
|| (res = executeNumRightType<T0, UInt16>(col_left, col_right_untyped))
|
|
|| (res = executeNumRightType<T0, UInt32>(col_left, col_right_untyped))
|
|
|| (res = executeNumRightType<T0, UInt64>(col_left, col_right_untyped))
|
|
|| (res = executeNumRightType<T0, UInt128>(col_left, col_right_untyped))
|
|
|| (res = executeNumRightType<T0, UInt256>(col_left, col_right_untyped))
|
|
|| (res = executeNumRightType<T0, Int8>(col_left, col_right_untyped))
|
|
|| (res = executeNumRightType<T0, Int16>(col_left, col_right_untyped))
|
|
|| (res = executeNumRightType<T0, Int32>(col_left, col_right_untyped))
|
|
|| (res = executeNumRightType<T0, Int64>(col_left, col_right_untyped))
|
|
|| (res = executeNumRightType<T0, Int128>(col_left, col_right_untyped))
|
|
|| (res = executeNumRightType<T0, Int256>(col_left, col_right_untyped))
|
|
|| (res = executeNumRightType<T0, Float32>(col_left, col_right_untyped))
|
|
|| (res = executeNumRightType<T0, Float64>(col_left, col_right_untyped)))
|
|
return res;
|
|
else
|
|
throw Exception("Illegal column " + col_right_untyped->getName()
|
|
+ " of second argument of function " + getName(),
|
|
ErrorCodes::ILLEGAL_COLUMN);
|
|
}
|
|
else if (auto col_left_const = checkAndGetColumnConst<ColumnVector<T0>>(col_left_untyped))
|
|
{
|
|
if ( (res = executeNumConstRightType<T0, UInt8>(col_left_const, col_right_untyped))
|
|
|| (res = executeNumConstRightType<T0, UInt16>(col_left_const, col_right_untyped))
|
|
|| (res = executeNumConstRightType<T0, UInt32>(col_left_const, col_right_untyped))
|
|
|| (res = executeNumConstRightType<T0, UInt64>(col_left_const, col_right_untyped))
|
|
|| (res = executeNumConstRightType<T0, UInt128>(col_left_const, col_right_untyped))
|
|
|| (res = executeNumConstRightType<T0, UInt256>(col_left_const, col_right_untyped))
|
|
|| (res = executeNumConstRightType<T0, Int8>(col_left_const, col_right_untyped))
|
|
|| (res = executeNumConstRightType<T0, Int16>(col_left_const, col_right_untyped))
|
|
|| (res = executeNumConstRightType<T0, Int32>(col_left_const, col_right_untyped))
|
|
|| (res = executeNumConstRightType<T0, Int64>(col_left_const, col_right_untyped))
|
|
|| (res = executeNumConstRightType<T0, Int128>(col_left_const, col_right_untyped))
|
|
|| (res = executeNumConstRightType<T0, Int256>(col_left_const, col_right_untyped))
|
|
|| (res = executeNumConstRightType<T0, Float32>(col_left_const, col_right_untyped))
|
|
|| (res = executeNumConstRightType<T0, Float64>(col_left_const, col_right_untyped)))
|
|
return res;
|
|
else
|
|
throw Exception("Illegal column " + col_right_untyped->getName()
|
|
+ " of second argument of function " + getName(),
|
|
ErrorCodes::ILLEGAL_COLUMN);
|
|
}
|
|
|
|
return nullptr;
|
|
}
|
|
|
|
ColumnPtr executeDecimal(const ColumnWithTypeAndName & col_left, const ColumnWithTypeAndName & col_right) const
|
|
{
|
|
TypeIndex left_number = col_left.type->getTypeId();
|
|
TypeIndex right_number = col_right.type->getTypeId();
|
|
ColumnPtr res;
|
|
|
|
auto call = [&](const auto & types) -> bool
|
|
{
|
|
using Types = std::decay_t<decltype(types)>;
|
|
using LeftDataType = typename Types::LeftType;
|
|
using RightDataType = typename Types::RightType;
|
|
|
|
if (check_decimal_overflow)
|
|
return (res = DecimalComparison<LeftDataType, RightDataType, Op, true>::apply(col_left, col_right)) != nullptr;
|
|
else
|
|
return (res = DecimalComparison<LeftDataType, RightDataType, Op, false>::apply(col_left, col_right)) != nullptr;
|
|
};
|
|
|
|
if (!callOnBasicTypes<true, false, true, true>(left_number, right_number, call))
|
|
throw Exception("Wrong call for " + getName() + " with " + col_left.type->getName() + " and " + col_right.type->getName(),
|
|
ErrorCodes::LOGICAL_ERROR);
|
|
|
|
return res;
|
|
}
|
|
|
|
ColumnPtr executeString(const IColumn * c0, const IColumn * c1) const
|
|
{
|
|
const ColumnString * c0_string = checkAndGetColumn<ColumnString>(c0);
|
|
const ColumnString * c1_string = checkAndGetColumn<ColumnString>(c1);
|
|
const ColumnFixedString * c0_fixed_string = checkAndGetColumn<ColumnFixedString>(c0);
|
|
const ColumnFixedString * c1_fixed_string = checkAndGetColumn<ColumnFixedString>(c1);
|
|
|
|
const ColumnConst * c0_const = checkAndGetColumnConstStringOrFixedString(c0);
|
|
const ColumnConst * c1_const = checkAndGetColumnConstStringOrFixedString(c1);
|
|
|
|
if (!((c0_string || c0_fixed_string || c0_const) && (c1_string || c1_fixed_string || c1_const)))
|
|
return nullptr;
|
|
|
|
const ColumnString::Chars * c0_const_chars = nullptr;
|
|
const ColumnString::Chars * c1_const_chars = nullptr;
|
|
ColumnString::Offset c0_const_size = 0;
|
|
ColumnString::Offset c1_const_size = 0;
|
|
|
|
if (c0_const)
|
|
{
|
|
const ColumnString * c0_const_string = checkAndGetColumn<ColumnString>(&c0_const->getDataColumn());
|
|
const ColumnFixedString * c0_const_fixed_string = checkAndGetColumn<ColumnFixedString>(&c0_const->getDataColumn());
|
|
|
|
if (c0_const_string)
|
|
{
|
|
c0_const_chars = &c0_const_string->getChars();
|
|
c0_const_size = c0_const_string->getDataAt(0).size;
|
|
}
|
|
else if (c0_const_fixed_string)
|
|
{
|
|
c0_const_chars = &c0_const_fixed_string->getChars();
|
|
c0_const_size = c0_const_fixed_string->getN();
|
|
}
|
|
else
|
|
throw Exception("Logical error: ColumnConst contains not String nor FixedString column", ErrorCodes::ILLEGAL_COLUMN);
|
|
}
|
|
|
|
if (c1_const)
|
|
{
|
|
const ColumnString * c1_const_string = checkAndGetColumn<ColumnString>(&c1_const->getDataColumn());
|
|
const ColumnFixedString * c1_const_fixed_string = checkAndGetColumn<ColumnFixedString>(&c1_const->getDataColumn());
|
|
|
|
if (c1_const_string)
|
|
{
|
|
c1_const_chars = &c1_const_string->getChars();
|
|
c1_const_size = c1_const_string->getDataAt(0).size;
|
|
}
|
|
else if (c1_const_fixed_string)
|
|
{
|
|
c1_const_chars = &c1_const_fixed_string->getChars();
|
|
c1_const_size = c1_const_fixed_string->getN();
|
|
}
|
|
else
|
|
throw Exception("Logical error: ColumnConst contains not String nor FixedString column", ErrorCodes::ILLEGAL_COLUMN);
|
|
}
|
|
|
|
using StringImpl = StringComparisonImpl<Op<int, int>>;
|
|
|
|
if (c0_const && c1_const)
|
|
{
|
|
auto res = executeString(&c0_const->getDataColumn(), &c1_const->getDataColumn());
|
|
if (!res)
|
|
return nullptr;
|
|
|
|
return ColumnConst::create(res, c0_const->size());
|
|
}
|
|
else
|
|
{
|
|
auto c_res = ColumnUInt8::create();
|
|
ColumnUInt8::Container & vec_res = c_res->getData();
|
|
vec_res.resize(c0->size());
|
|
|
|
if (c0_string && c1_string)
|
|
StringImpl::string_vector_string_vector(
|
|
c0_string->getChars(), c0_string->getOffsets(),
|
|
c1_string->getChars(), c1_string->getOffsets(),
|
|
c_res->getData());
|
|
else if (c0_string && c1_fixed_string)
|
|
StringImpl::string_vector_fixed_string_vector(
|
|
c0_string->getChars(), c0_string->getOffsets(),
|
|
c1_fixed_string->getChars(), c1_fixed_string->getN(),
|
|
c_res->getData());
|
|
else if (c0_string && c1_const)
|
|
StringImpl::string_vector_constant(
|
|
c0_string->getChars(), c0_string->getOffsets(),
|
|
*c1_const_chars, c1_const_size,
|
|
c_res->getData());
|
|
else if (c0_fixed_string && c1_string)
|
|
StringImpl::fixed_string_vector_string_vector(
|
|
c0_fixed_string->getChars(), c0_fixed_string->getN(),
|
|
c1_string->getChars(), c1_string->getOffsets(),
|
|
c_res->getData());
|
|
else if (c0_fixed_string && c1_fixed_string)
|
|
StringImpl::fixed_string_vector_fixed_string_vector(
|
|
c0_fixed_string->getChars(), c0_fixed_string->getN(),
|
|
c1_fixed_string->getChars(), c1_fixed_string->getN(),
|
|
c_res->getData());
|
|
else if (c0_fixed_string && c1_const)
|
|
StringImpl::fixed_string_vector_constant(
|
|
c0_fixed_string->getChars(), c0_fixed_string->getN(),
|
|
*c1_const_chars, c1_const_size,
|
|
c_res->getData());
|
|
else if (c0_const && c1_string)
|
|
StringImpl::constant_string_vector(
|
|
*c0_const_chars, c0_const_size,
|
|
c1_string->getChars(), c1_string->getOffsets(),
|
|
c_res->getData());
|
|
else if (c0_const && c1_fixed_string)
|
|
StringImpl::constant_fixed_string_vector(
|
|
*c0_const_chars, c0_const_size,
|
|
c1_fixed_string->getChars(), c1_fixed_string->getN(),
|
|
c_res->getData());
|
|
else
|
|
throw Exception("Illegal columns "
|
|
+ c0->getName() + " and " + c1->getName()
|
|
+ " of arguments of function " + getName(),
|
|
ErrorCodes::ILLEGAL_COLUMN);
|
|
|
|
return c_res;
|
|
}
|
|
}
|
|
|
|
ColumnPtr executeWithConstString(
|
|
const DataTypePtr & result_type, const IColumn * col_left_untyped, const IColumn * col_right_untyped,
|
|
const DataTypePtr & left_type, const DataTypePtr & right_type, size_t input_rows_count) const
|
|
{
|
|
/// To compare something with const string, we cast constant to appropriate type and compare as usual.
|
|
/// It is ok to throw exception if value is not convertible.
|
|
/// We should deal with possible overflows, e.g. toUInt8(1) = '257' should return false.
|
|
|
|
const ColumnConst * left_const = checkAndGetColumnConstStringOrFixedString(col_left_untyped);
|
|
const ColumnConst * right_const = checkAndGetColumnConstStringOrFixedString(col_right_untyped);
|
|
|
|
if (!left_const && !right_const)
|
|
return nullptr;
|
|
|
|
const IDataType * type_string = left_const ? left_type.get() : right_type.get();
|
|
const DataTypePtr & type_to_compare = !left_const ? left_type : right_type;
|
|
|
|
Field string_value = left_const ? left_const->getField() : right_const->getField();
|
|
Field converted = convertFieldToType(string_value, *type_to_compare, type_string);
|
|
|
|
/// If not possible to convert, comparison with =, <, >, <=, >= yields to false and comparison with != yields to true.
|
|
if (converted.isNull())
|
|
{
|
|
return DataTypeUInt8().createColumnConst(input_rows_count, IsOperation<Op>::not_equals);
|
|
}
|
|
else
|
|
{
|
|
auto column_converted = type_to_compare->createColumnConst(input_rows_count, converted);
|
|
|
|
ColumnsWithTypeAndName tmp_columns
|
|
{
|
|
{ left_const ? column_converted : col_left_untyped->getPtr(), type_to_compare, "" },
|
|
{ !left_const ? column_converted : col_right_untyped->getPtr(), type_to_compare, "" },
|
|
};
|
|
|
|
return executeImpl(tmp_columns, result_type, input_rows_count);
|
|
}
|
|
}
|
|
|
|
ColumnPtr executeTuple(
|
|
const DataTypePtr & result_type, const ColumnWithTypeAndName & c0, const ColumnWithTypeAndName & c1,
|
|
size_t input_rows_count) const
|
|
{
|
|
/** We will lexicographically compare the tuples. This is done as follows:
|
|
* x == y : x1 == y1 && x2 == y2 ...
|
|
* x != y : x1 != y1 || x2 != y2 ...
|
|
*
|
|
* x < y: x1 < y1 || (x1 == y1 && (x2 < y2 || (x2 == y2 ... && xn < yn))
|
|
* x > y: x1 > y1 || (x1 == y1 && (x2 > y2 || (x2 == y2 ... && xn > yn))
|
|
* x <= y: x1 < y1 || (x1 == y1 && (x2 < y2 || (x2 == y2 ... && xn <= yn))
|
|
*
|
|
* Recursive form:
|
|
* x <= y: x1 < y1 || (x1 == y1 && x_tail <= y_tail)
|
|
*
|
|
* x >= y: x1 > y1 || (x1 == y1 && (x2 > y2 || (x2 == y2 ... && xn >= yn))
|
|
*/
|
|
|
|
const size_t tuple_size = typeid_cast<const DataTypeTuple &>(*c0.type).getElements().size();
|
|
|
|
if (0 == tuple_size)
|
|
throw Exception("Comparison of zero-sized tuples is not implemented.", ErrorCodes::NOT_IMPLEMENTED);
|
|
|
|
if (tuple_size != typeid_cast<const DataTypeTuple &>(*c1.type).getElements().size())
|
|
throw Exception("Cannot compare tuples of different sizes.", ErrorCodes::BAD_ARGUMENTS);
|
|
|
|
if (result_type->onlyNull())
|
|
return result_type->createColumnConstWithDefaultValue(input_rows_count);
|
|
|
|
ColumnsWithTypeAndName x(tuple_size);
|
|
ColumnsWithTypeAndName y(tuple_size);
|
|
|
|
const auto * x_const = checkAndGetColumnConst<ColumnTuple>(c0.column.get());
|
|
const auto * y_const = checkAndGetColumnConst<ColumnTuple>(c1.column.get());
|
|
|
|
Columns x_columns;
|
|
Columns y_columns;
|
|
|
|
if (x_const)
|
|
x_columns = convertConstTupleToConstantElements(*x_const);
|
|
else
|
|
x_columns = assert_cast<const ColumnTuple &>(*c0.column).getColumnsCopy();
|
|
|
|
if (y_const)
|
|
y_columns = convertConstTupleToConstantElements(*y_const);
|
|
else
|
|
y_columns = assert_cast<const ColumnTuple &>(*c1.column).getColumnsCopy();
|
|
|
|
for (size_t i = 0; i < tuple_size; ++i)
|
|
{
|
|
x[i].type = static_cast<const DataTypeTuple &>(*c0.type).getElements()[i];
|
|
y[i].type = static_cast<const DataTypeTuple &>(*c1.type).getElements()[i];
|
|
|
|
x[i].column = x_columns[i];
|
|
y[i].column = y_columns[i];
|
|
}
|
|
|
|
return executeTupleImpl(x, y, tuple_size, input_rows_count);
|
|
}
|
|
|
|
ColumnPtr executeTupleImpl(const ColumnsWithTypeAndName & x,
|
|
const ColumnsWithTypeAndName & y, size_t tuple_size,
|
|
size_t input_rows_count) const;
|
|
|
|
ColumnPtr executeTupleEqualityImpl(
|
|
std::shared_ptr<IFunctionOverloadResolver> func_compare,
|
|
std::shared_ptr<IFunctionOverloadResolver> func_convolution,
|
|
const ColumnsWithTypeAndName & x,
|
|
const ColumnsWithTypeAndName & y,
|
|
size_t tuple_size,
|
|
size_t input_rows_count) const
|
|
{
|
|
if (0 == tuple_size)
|
|
throw Exception("Comparison of zero-sized tuples is not implemented.", ErrorCodes::NOT_IMPLEMENTED);
|
|
|
|
ColumnsWithTypeAndName convolution_columns(tuple_size);
|
|
ColumnsWithTypeAndName tmp_columns(2);
|
|
|
|
for (size_t i = 0; i < tuple_size; ++i)
|
|
{
|
|
tmp_columns[0] = x[i];
|
|
tmp_columns[1] = y[i];
|
|
|
|
auto impl = func_compare->build(tmp_columns);
|
|
convolution_columns[i].type = impl->getResultType();
|
|
|
|
/// Comparison of the elements.
|
|
convolution_columns[i].column = impl->execute(tmp_columns, impl->getResultType(), input_rows_count);
|
|
}
|
|
|
|
if (tuple_size == 1)
|
|
{
|
|
/// Do not call AND for single-element tuple.
|
|
return convolution_columns[0].column;
|
|
}
|
|
|
|
/// Logical convolution.
|
|
auto impl = func_convolution->build(convolution_columns);
|
|
return impl->execute(convolution_columns, impl->getResultType(), input_rows_count);
|
|
}
|
|
|
|
ColumnPtr executeTupleLessGreaterImpl(
|
|
std::shared_ptr<IFunctionOverloadResolver> func_compare_head,
|
|
std::shared_ptr<IFunctionOverloadResolver> func_compare_tail,
|
|
std::shared_ptr<IFunctionOverloadResolver> func_and,
|
|
std::shared_ptr<IFunctionOverloadResolver> func_or,
|
|
std::shared_ptr<IFunctionOverloadResolver> func_equals,
|
|
const ColumnsWithTypeAndName & x,
|
|
const ColumnsWithTypeAndName & y,
|
|
size_t tuple_size,
|
|
size_t input_rows_count) const
|
|
{
|
|
ColumnsWithTypeAndName less_columns(tuple_size);
|
|
ColumnsWithTypeAndName equal_columns(tuple_size - 1);
|
|
ColumnsWithTypeAndName tmp_columns(2);
|
|
|
|
/// Pairwise comparison of the inequality of all elements; on the equality of all elements except the last.
|
|
/// (x[i], y[i], x[i] < y[i], x[i] == y[i])
|
|
for (size_t i = 0; i < tuple_size; ++i)
|
|
{
|
|
tmp_columns[0] = x[i];
|
|
tmp_columns[1] = y[i];
|
|
|
|
if (i + 1 != tuple_size)
|
|
{
|
|
auto impl_head = func_compare_head->build(tmp_columns);
|
|
less_columns[i].type = impl_head->getResultType();
|
|
less_columns[i].column = impl_head->execute(tmp_columns, less_columns[i].type, input_rows_count);
|
|
|
|
auto impl_equals = func_equals->build(tmp_columns);
|
|
equal_columns[i].type = impl_equals->getResultType();
|
|
equal_columns[i].column = impl_equals->execute(tmp_columns, equal_columns[i].type, input_rows_count);
|
|
|
|
}
|
|
else
|
|
{
|
|
auto impl_tail = func_compare_tail->build(tmp_columns);
|
|
less_columns[i].type = impl_tail->getResultType();
|
|
less_columns[i].column = impl_tail->execute(tmp_columns, less_columns[i].type, input_rows_count);
|
|
}
|
|
}
|
|
|
|
/// Combination. Complex code - make a drawing. It can be replaced by a recursive comparison of tuples.
|
|
/// Last column contains intermediate result.
|
|
/// Code is generally equivalent to:
|
|
/// res = `x < y`[tuple_size - 1];
|
|
/// for (int i = tuple_size - 2; i >= 0; --i)
|
|
/// res = (res && `x == y`[i]) || `x < y`[i];
|
|
size_t i = tuple_size - 1;
|
|
tmp_columns[0] = less_columns[i];
|
|
while (i > 0)
|
|
{
|
|
--i;
|
|
|
|
tmp_columns[1] = equal_columns[i];
|
|
auto func_and_adaptor = func_and->build(tmp_columns);
|
|
|
|
tmp_columns[0].column = func_and_adaptor->execute(tmp_columns, func_and_adaptor->getResultType(), input_rows_count);
|
|
tmp_columns[0].type = func_and_adaptor->getResultType();
|
|
|
|
tmp_columns[1] = less_columns[i];
|
|
auto func_or_adaptor = func_or->build(tmp_columns);
|
|
|
|
tmp_columns[0].column = func_or_adaptor->execute(tmp_columns, func_or_adaptor->getResultType(), input_rows_count);
|
|
tmp_columns[tmp_columns.size() - 1].type = func_or_adaptor->getResultType();
|
|
}
|
|
|
|
return tmp_columns[0].column;
|
|
}
|
|
|
|
ColumnPtr executeGenericIdenticalTypes(const IColumn * c0, const IColumn * c1) const
|
|
{
|
|
bool c0_const = isColumnConst(*c0);
|
|
bool c1_const = isColumnConst(*c1);
|
|
|
|
if (c0_const && c1_const)
|
|
{
|
|
UInt8 res = 0;
|
|
GenericComparisonImpl<Op<int, int>>::constantConstant(*c0, *c1, res);
|
|
return DataTypeUInt8().createColumnConst(c0->size(), toField(res));
|
|
}
|
|
else
|
|
{
|
|
auto c_res = ColumnUInt8::create();
|
|
ColumnUInt8::Container & vec_res = c_res->getData();
|
|
vec_res.resize(c0->size());
|
|
|
|
if (c0_const)
|
|
GenericComparisonImpl<Op<int, int>>::constantVector(*c0, *c1, vec_res);
|
|
else if (c1_const)
|
|
GenericComparisonImpl<Op<int, int>>::vectorConstant(*c0, *c1, vec_res);
|
|
else
|
|
GenericComparisonImpl<Op<int, int>>::vectorVector(*c0, *c1, vec_res);
|
|
|
|
return c_res;
|
|
}
|
|
}
|
|
|
|
ColumnPtr executeGeneric(const ColumnWithTypeAndName & c0, const ColumnWithTypeAndName & c1) const
|
|
{
|
|
DataTypePtr common_type = getLeastSupertype(DataTypes{c0.type, c1.type});
|
|
|
|
ColumnPtr c0_converted = castColumn(c0, common_type);
|
|
ColumnPtr c1_converted = castColumn(c1, common_type);
|
|
|
|
return executeGenericIdenticalTypes(c0_converted.get(), c1_converted.get());
|
|
}
|
|
|
|
public:
|
|
String getName() const override
|
|
{
|
|
return name;
|
|
}
|
|
|
|
size_t getNumberOfArguments() const override { return 2; }
|
|
|
|
bool isSuitableForShortCircuitArgumentsExecution(const DataTypesWithConstInfo & /*arguments*/) const override { return false; }
|
|
|
|
/// Get result types by argument types. If the function does not apply to these arguments, throw an exception.
|
|
DataTypePtr getReturnTypeImpl(const DataTypes & arguments) const override
|
|
{
|
|
WhichDataType left(arguments[0].get());
|
|
WhichDataType right(arguments[1].get());
|
|
|
|
const DataTypeTuple * left_tuple = checkAndGetDataType<DataTypeTuple>(arguments[0].get());
|
|
const DataTypeTuple * right_tuple = checkAndGetDataType<DataTypeTuple>(arguments[1].get());
|
|
|
|
bool both_represented_by_number = arguments[0]->isValueRepresentedByNumber() && arguments[1]->isValueRepresentedByNumber();
|
|
bool has_date = left.isDateOrDate32() || right.isDateOrDate32();
|
|
|
|
if (!((both_represented_by_number && !has_date) /// Do not allow to compare date and number.
|
|
|| (left.isStringOrFixedString() || right.isStringOrFixedString()) /// Everything can be compared with string by conversion.
|
|
/// You can compare the date, datetime, or datatime64 and an enumeration with a constant string.
|
|
|| ((left.isDate() || left.isDate32() || left.isDateTime() || left.isDateTime64()) && (right.isDate() || right.isDate32() || right.isDateTime() || right.isDateTime64()) && left.idx == right.idx) /// only date vs date, or datetime vs datetime
|
|
|| (left.isUUID() && right.isUUID())
|
|
|| (left.isEnum() && right.isEnum() && arguments[0]->getName() == arguments[1]->getName()) /// only equivalent enum type values can be compared against
|
|
|| (left_tuple && right_tuple && left_tuple->getElements().size() == right_tuple->getElements().size())
|
|
|| (arguments[0]->equals(*arguments[1]))))
|
|
{
|
|
try
|
|
{
|
|
getLeastSupertype(arguments);
|
|
}
|
|
catch (const Exception &)
|
|
{
|
|
throw Exception("Illegal types of arguments (" + arguments[0]->getName() + ", " + arguments[1]->getName() + ")"
|
|
" of function " + getName(), ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT);
|
|
}
|
|
}
|
|
|
|
if (left_tuple && right_tuple)
|
|
{
|
|
auto func = FunctionToOverloadResolverAdaptor(FunctionComparison<Op, Name>::create(context));
|
|
|
|
bool has_nullable = false;
|
|
bool has_null = false;
|
|
|
|
size_t size = left_tuple->getElements().size();
|
|
for (size_t i = 0; i < size; ++i)
|
|
{
|
|
ColumnsWithTypeAndName args = {{nullptr, left_tuple->getElements()[i], ""},
|
|
{nullptr, right_tuple->getElements()[i], ""}};
|
|
auto element_type = func.build(args)->getResultType();
|
|
has_nullable = has_nullable || element_type->isNullable();
|
|
has_null = has_null || element_type->onlyNull();
|
|
}
|
|
|
|
/// If any element comparison is nullable, return type will also be nullable.
|
|
/// We useDefaultImplementationForNulls, but it doesn't work for tuples.
|
|
if (has_null)
|
|
return std::make_shared<DataTypeNullable>(std::make_shared<DataTypeNothing>());
|
|
if (has_nullable)
|
|
return std::make_shared<DataTypeNullable>(std::make_shared<DataTypeUInt8>());
|
|
}
|
|
|
|
return std::make_shared<DataTypeUInt8>();
|
|
}
|
|
|
|
ColumnPtr executeImpl(const ColumnsWithTypeAndName & arguments, const DataTypePtr & result_type, size_t input_rows_count) const override
|
|
{
|
|
const auto & col_with_type_and_name_left = arguments[0];
|
|
const auto & col_with_type_and_name_right = arguments[1];
|
|
const IColumn * col_left_untyped = col_with_type_and_name_left.column.get();
|
|
const IColumn * col_right_untyped = col_with_type_and_name_right.column.get();
|
|
|
|
const DataTypePtr & left_type = col_with_type_and_name_left.type;
|
|
const DataTypePtr & right_type = col_with_type_and_name_right.type;
|
|
|
|
/// The case when arguments are the same (tautological comparison). Return constant.
|
|
/// NOTE: Nullable types are special case.
|
|
/// (BTW, this function use default implementation for Nullable, so Nullable types cannot be here. Check just in case.)
|
|
/// NOTE: We consider NaN comparison to be implementation specific (and in our implementation NaNs are sometimes equal sometimes not).
|
|
if (left_type->equals(*right_type) && !left_type->isNullable() && !isTuple(left_type) && col_left_untyped == col_right_untyped)
|
|
{
|
|
ColumnPtr result_column;
|
|
|
|
/// Always true: =, <=, >=
|
|
if constexpr (IsOperation<Op>::equals
|
|
|| IsOperation<Op>::less_or_equals
|
|
|| IsOperation<Op>::greater_or_equals)
|
|
{
|
|
result_column = DataTypeUInt8().createColumnConst(input_rows_count, 1u);
|
|
}
|
|
else
|
|
{
|
|
result_column = DataTypeUInt8().createColumnConst(input_rows_count, 0u);
|
|
}
|
|
|
|
if (!isColumnConst(*col_left_untyped))
|
|
result_column = result_column->convertToFullColumnIfConst();
|
|
|
|
return result_column;
|
|
}
|
|
|
|
WhichDataType which_left{left_type};
|
|
WhichDataType which_right{right_type};
|
|
|
|
const bool left_is_num = col_left_untyped->isNumeric();
|
|
const bool right_is_num = col_right_untyped->isNumeric();
|
|
|
|
const bool left_is_string = which_left.isStringOrFixedString();
|
|
const bool right_is_string = which_right.isStringOrFixedString();
|
|
|
|
const bool left_is_float = which_left.isFloat();
|
|
const bool right_is_float = which_right.isFloat();
|
|
|
|
bool date_and_datetime = (which_left.idx != which_right.idx) && (which_left.isDate() || which_left.isDate32() || which_left.isDateTime() || which_left.isDateTime64())
|
|
&& (which_right.isDate() || which_right.isDate32() || which_right.isDateTime() || which_right.isDateTime64());
|
|
|
|
ColumnPtr res;
|
|
if (left_is_num && right_is_num && !date_and_datetime)
|
|
{
|
|
if (!((res = executeNumLeftType<UInt8>(col_left_untyped, col_right_untyped))
|
|
|| (res = executeNumLeftType<UInt16>(col_left_untyped, col_right_untyped))
|
|
|| (res = executeNumLeftType<UInt32>(col_left_untyped, col_right_untyped))
|
|
|| (res = executeNumLeftType<UInt64>(col_left_untyped, col_right_untyped))
|
|
|| (res = executeNumLeftType<UInt128>(col_left_untyped, col_right_untyped))
|
|
|| (res = executeNumLeftType<UInt256>(col_left_untyped, col_right_untyped))
|
|
|| (res = executeNumLeftType<Int8>(col_left_untyped, col_right_untyped))
|
|
|| (res = executeNumLeftType<Int16>(col_left_untyped, col_right_untyped))
|
|
|| (res = executeNumLeftType<Int32>(col_left_untyped, col_right_untyped))
|
|
|| (res = executeNumLeftType<Int64>(col_left_untyped, col_right_untyped))
|
|
|| (res = executeNumLeftType<Int128>(col_left_untyped, col_right_untyped))
|
|
|| (res = executeNumLeftType<Int256>(col_left_untyped, col_right_untyped))
|
|
|| (res = executeNumLeftType<Float32>(col_left_untyped, col_right_untyped))
|
|
|| (res = executeNumLeftType<Float64>(col_left_untyped, col_right_untyped))))
|
|
throw Exception("Illegal column " + col_left_untyped->getName()
|
|
+ " of first argument of function " + getName(),
|
|
ErrorCodes::ILLEGAL_COLUMN);
|
|
|
|
return res;
|
|
}
|
|
else if (checkAndGetDataType<DataTypeTuple>(left_type.get())
|
|
&& checkAndGetDataType<DataTypeTuple>(right_type.get()))
|
|
{
|
|
return executeTuple(result_type, col_with_type_and_name_left, col_with_type_and_name_right, input_rows_count);
|
|
}
|
|
else if (left_is_string && right_is_string && (res = executeString(col_left_untyped, col_right_untyped)))
|
|
{
|
|
return res;
|
|
}
|
|
else if ((res = executeWithConstString(
|
|
result_type, col_left_untyped, col_right_untyped,
|
|
left_type, right_type,
|
|
input_rows_count)))
|
|
{
|
|
return res;
|
|
}
|
|
else if ((isColumnedAsDecimal(left_type) || isColumnedAsDecimal(right_type)))
|
|
{
|
|
// Comparing Date/Date32 and DateTime64 requires implicit conversion,
|
|
if (date_and_datetime && (isDateOrDate32(left_type) || isDateOrDate32(right_type)))
|
|
{
|
|
DataTypePtr common_type = getLeastSupertype(DataTypes{left_type, right_type});
|
|
ColumnPtr c0_converted = castColumn(col_with_type_and_name_left, common_type);
|
|
ColumnPtr c1_converted = castColumn(col_with_type_and_name_right, common_type);
|
|
return executeDecimal({c0_converted, common_type, "left"}, {c1_converted, common_type, "right"});
|
|
}
|
|
else
|
|
{
|
|
/// Check does another data type is comparable to Decimal, includes Int and Float.
|
|
if (!allowDecimalComparison(left_type, right_type) && !date_and_datetime)
|
|
throw Exception(
|
|
"No operation " + getName() + " between " + left_type->getName() + " and " + right_type->getName(),
|
|
ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT);
|
|
/// When Decimal comparing to Float32/64, we convert both of them into Float64.
|
|
/// Other systems like MySQL and Spark also do as this.
|
|
if (left_is_float || right_is_float)
|
|
{
|
|
const auto converted_type = std::make_shared<DataTypeFloat64>();
|
|
ColumnPtr c0_converted = castColumn(col_with_type_and_name_left, converted_type);
|
|
ColumnPtr c1_converted = castColumn(col_with_type_and_name_right, converted_type);
|
|
|
|
auto new_arguments
|
|
= ColumnsWithTypeAndName{{c0_converted, converted_type, "left"}, {c1_converted, converted_type, "right"}};
|
|
return executeImpl(new_arguments, result_type, input_rows_count);
|
|
}
|
|
return executeDecimal(col_with_type_and_name_left, col_with_type_and_name_right);
|
|
}
|
|
|
|
}
|
|
else if (date_and_datetime)
|
|
{
|
|
DataTypePtr common_type = getLeastSupertype(DataTypes{left_type, right_type});
|
|
ColumnPtr c0_converted = castColumn(col_with_type_and_name_left, common_type);
|
|
ColumnPtr c1_converted = castColumn(col_with_type_and_name_right, common_type);
|
|
if (!((res = executeNumLeftType<UInt32>(c0_converted.get(), c1_converted.get()))
|
|
|| (res = executeNumLeftType<UInt64>(c0_converted.get(), c1_converted.get()))
|
|
|| (res = executeNumLeftType<Int32>(c0_converted.get(), c1_converted.get()))
|
|
|| (res = executeDecimal({c0_converted, common_type, "left"}, {c1_converted, common_type, "right"}))))
|
|
throw Exception("Date related common types can only be UInt32/UInt64/Int32/Decimal", ErrorCodes::LOGICAL_ERROR);
|
|
return res;
|
|
}
|
|
else if (left_type->equals(*right_type))
|
|
{
|
|
return executeGenericIdenticalTypes(col_left_untyped, col_right_untyped);
|
|
}
|
|
else
|
|
{
|
|
return executeGeneric(col_with_type_and_name_left, col_with_type_and_name_right);
|
|
}
|
|
}
|
|
|
|
#if USE_EMBEDDED_COMPILER
|
|
bool isCompilableImpl(const DataTypes & types) const override
|
|
{
|
|
if (2 != types.size())
|
|
return false;
|
|
|
|
WhichDataType data_type_lhs(types[0]);
|
|
WhichDataType data_type_rhs(types[1]);
|
|
|
|
auto is_big_integer = [](WhichDataType type) { return type.isUInt64() || type.isInt64(); };
|
|
|
|
if ((is_big_integer(data_type_lhs) && data_type_rhs.isFloat())
|
|
|| (is_big_integer(data_type_rhs) && data_type_lhs.isFloat())
|
|
|| (data_type_lhs.isDate() && data_type_rhs.isDateTime())
|
|
|| (data_type_rhs.isDate() && data_type_lhs.isDateTime()))
|
|
return false; /// TODO: implement (double, int_N where N > double's mantissa width)
|
|
|
|
return isCompilableType(types[0]) && isCompilableType(types[1]);
|
|
}
|
|
|
|
llvm::Value * compileImpl(llvm::IRBuilderBase & builder, const DataTypes & types, Values values) const override
|
|
{
|
|
assert(2 == types.size() && 2 == values.size());
|
|
|
|
auto & b = static_cast<llvm::IRBuilder<> &>(builder);
|
|
auto [x, y] = nativeCastToCommon(b, types[0], values[0], types[1], values[1]);
|
|
auto * result = CompileOp<Op>::compile(b, x, y, typeIsSigned(*types[0]) || typeIsSigned(*types[1]));
|
|
return b.CreateSelect(result, b.getInt8(1), b.getInt8(0));
|
|
}
|
|
#endif
|
|
};
|
|
|
|
}
|