Merge pull request #68639 from m7kss1/ripedmd-160

Add RIPEMD160 function

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

@@ -688,6 +688,40 @@ SELECT kostikConsistentHash(16045690984833335023, 2);
 └───────────────────────────────────────────────┘
 ```
 
+## ripeMD160
+
+Produces [RIPEMD-160](https://en.wikipedia.org/wiki/RIPEMD) hash value.
+
+**Syntax**
+
+```sql
+ripeMD160(input)
+```
+
+**Parameters**
+
+- `input`: Input string. [String](../data-types/string.md)
+
+**Returned value**
+
+- A [UInt256](../data-types/int-uint.md) hash value where the 160-bit RIPEMD-160 hash is stored in the first 20 bytes. The remaining 12 bytes are zero-padded.
+
+**Example**
+
+Use the [hex](../functions/encoding-functions.md/#hex) function to represent the result as a hex-encoded string.
+
+Query:
+
+```sql
+SELECT hex(ripeMD160('The quick brown fox jumps over the lazy dog'));
+```
+
+```response
+┌─hex(ripeMD160('The quick brown fox jumps over the lazy dog'))─┐
+│ 37F332F68DB77BD9D7EDD4969571AD671CF9DD3B                      │
+└───────────────────────────────────────────────────────────────┘
+```
+
 ## murmurHash2_32, murmurHash2_64
 
 Produces a [MurmurHash2](https://github.com/aappleby/smhasher) hash value.

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

@@ -124,6 +124,40 @@ SELECT hex(sipHash128('foo', '\x01', 3));
 └──────────────────────────────────┘
 ```
 
+## ripeMD160
+
+Генерирует [RIPEMD-160](https://en.wikipedia.org/wiki/RIPEMD) хеш строки.
+
+**Синтаксис**
+
+```sql
+ripeMD160(input)
+```
+
+**Аргументы**
+
+- `input`: Строка [String](../data-types/string.md)
+
+**Возвращаемое значение**
+
+- [UInt256](../data-types/int-uint.md), где 160-битный хеш RIPEMD-160 хранится в первых 20 байтах. Оставшиеся 12 байт заполняются нулями.
+
+**Пример**
+
+Используйте функцию [hex](../functions/encoding-functions.md#hex) для представления результата в виде строки с шестнадцатеричной кодировкой
+
+Запрос:
+
+```sql
+SELECT hex(ripeMD160('The quick brown fox jumps over the lazy dog'));
+```
+Результат:
+```response
+┌─hex(ripeMD160('The quick brown fox jumps over the lazy dog'))─┐
+│ 37F332F68DB77BD9D7EDD4969571AD671CF9DD3B                      │
+└───────────────────────────────────────────────────────────────┘
+```
+
 ## cityHash64 {#cityhash64}
 
 Генерирует 64-х битное значение [CityHash](https://github.com/google/cityhash).

src/Functions/FunctionsHashing.h

@@ -19,7 +19,9 @@
 #include <Common/HashTable/Hash.h>
 
 #if USE_SSL
+#    include <openssl/evp.h>
 #    include <openssl/md5.h>
+#    include <openssl/ripemd.h>
 #endif
 
 #include <bit>
@@ -196,6 +198,34 @@ T combineHashesFunc(T t1, T t2)
     return HashFunction::apply(reinterpret_cast<const char *>(hashes), sizeof(hashes));
 }
 
+#if USE_SSL
+struct RipeMD160Impl
+{
+    static constexpr auto name = "ripeMD160";
+    using ReturnType = UInt256;
+
+    static UInt256 apply(const char * begin, size_t size)
+    {
+        UInt8 digest[RIPEMD160_DIGEST_LENGTH];
+
+        RIPEMD160(reinterpret_cast<const unsigned char *>(begin), size, reinterpret_cast<unsigned char *>(digest));
+
+        std::reverse(digest, digest + RIPEMD160_DIGEST_LENGTH);
+
+        UInt256 res = 0;
+        std::memcpy(&res, digest, RIPEMD160_DIGEST_LENGTH);
+
+        return res;
+    }
+
+    static UInt256 combineHashes(UInt256 h1, UInt256 h2)
+    {
+        return combineHashesFunc<UInt256, RipeMD160Impl>(h1, h2);
+    }
+
+    static constexpr bool use_int_hash_for_pods = false;
+};
+#endif
 
 struct SipHash64Impl
 {
@@ -1624,6 +1654,7 @@ using FunctionIntHash32 = FunctionIntHash<IntHash32Impl, NameIntHash32>;
 using FunctionIntHash64 = FunctionIntHash<IntHash64Impl, NameIntHash64>;
 #if USE_SSL
 using FunctionHalfMD5 = FunctionAnyHash<HalfMD5Impl>;
+using FunctionRipeMD160Hash = FunctionAnyHash<RipeMD160Impl>;
 #endif
 using FunctionSipHash128 = FunctionAnyHash<SipHash128Impl>;
 using FunctionSipHash128Keyed = FunctionAnyHash<SipHash128KeyedImpl, true, SipHash128KeyedImpl::Key, SipHash128KeyedImpl::KeyColumns>;
@@ -1652,6 +1683,7 @@ using FunctionXxHash64 = FunctionAnyHash<ImplXxHash64>;
 using FunctionXXH3 = FunctionAnyHash<ImplXXH3>;
 
 using FunctionWyHash64 = FunctionAnyHash<ImplWyHash64>;
+
 }
 
 #pragma clang diagnostic pop

src/Functions/FunctionsHashingRipe.cpp

@@ -0,0 +1,23 @@
+#include "FunctionsHashing.h"
+
+#include <Functions/FunctionFactory.h>
+
+/// FunctionsHashing instantiations are separated into files FunctionsHashing*.cpp
+/// to better parallelize the build procedure and avoid MSan build failure
+/// due to excessive resource consumption.
+namespace DB
+{
+#if USE_SSL
+REGISTER_FUNCTION(HashingRipe)
+{
+    factory.registerFunction<FunctionRipeMD160Hash>(FunctionDocumentation{
+        .description = "RIPEMD-160 hash function, primarily used in Bitcoin address generation.",
+        .examples{{"", "SELECT hex(ripeMD160('The quick brown fox jumps over the lazy dog'));", R"(
+            ┌─hex(ripeMD160('The quick brown fox jumps over the lazy dog'))─┐
+            │ 37F332F68DB77BD9D7EDD4969571AD671CF9DD3B                      │
+            └───────────────────────────────────────────────────────────────┘
+        )"}},
+        .categories{"Hash"}});
+}
+#endif
+}

tests/queries/0_stateless/03222_ripeMD160.reference

@@ -0,0 +1,5 @@
+37F332F68DB77BD9D7EDD4969571AD671CF9DD3B
+132072DF690933835EB8B6AD0B77E7B6F14ACAD7
+9C1185A5C5E9FC54612808977EE8F548B2258D31
+13920F39C93D503A0AC02EAB9AA8F672BC523ADA
+3FEDF0C212CCFA54C0EBA676C8A8A2A10BC218BE

tests/queries/0_stateless/03222_ripeMD160.sql

@@ -0,0 +1,12 @@
+-- Tags: no-fasttest
+-- Ouput can be verified using: https://emn178.github.io/online-tools/ripemd-160/
+
+SELECT hex(ripeMD160('The quick brown fox jumps over the lazy dog'));
+
+SELECT hex(ripeMD160('The quick brown fox jumps over the lazy cog'));
+
+SELECT hex(ripeMD160(''));
+
+SELECT hex(ripeMD160('CheREpaha1512'));
+
+SELECT hex(ripeMD160('A-very-long-string-that-should-be-hashed-using-ripeMD160'));

utils/check-style/aspell-ignore/en/aspell-dict.txt

@@ -1,4 +1,4 @@
-personal_ws-1.1 en 2942
+personal_ws-1.1 en 2983
 AArch
 ACLs
 ALTERs
@@ -975,6 +975,7 @@ ThreadPoolRemoteFSReaderThreads
 ThreadPoolRemoteFSReaderThreadsActive
 ThreadsActive
 ThreadsInOvercommitTracker
+TimeSeries
 Timeunit
 TinyLog
 Tkachenko
@@ -1116,12 +1117,12 @@ addressToLineWithInlines
 addressToSymbol
 adviced
 agg
+aggThrow
 aggregatefunction
 aggregatingmergetree
 aggregatio
 aggretate
 aggthrow
-aggThrow
 aiochclient
 allocator
 alphaTokens
@@ -1895,8 +1896,8 @@ joinGet
 joinGetOrNull
 json
 jsonMergePatch
-jsonasstring
 jsonasobject
+jsonasstring
 jsoncolumns
 jsoncolumnsmonoblock
 jsoncompact
@@ -1937,8 +1938,8 @@ kurtSamp
 kurtosis
 kurtpop
 kurtsamp
-laion
 lagInFrame
+laion
 lang
 laravel
 largestTriangleThreeBuckets
@@ -2040,7 +2041,6 @@ maxMap
 maxintersections
 maxintersectionsposition
 maxmap
-minMappedArrays
 maxmind
 mdadm
 meanZTest
@@ -2237,8 +2237,8 @@ parseReadableSizeOrZero
 parseTimeDelta
 parseable
 parsers
-partitionId
 partitionID
+partitionId
 pathFull
 pclmulqdq
 pcre
@@ -2467,6 +2467,7 @@ rewritable
 rightPad
 rightPadUTF
 rightUTF
+ripeMD
 risc
 riscv
 ro
@@ -2718,7 +2719,6 @@ themself
 threadpool
 throwIf
 timeDiff
-TimeSeries
 timeSeriesData
 timeSeriesMetrics
 timeSeriesTags