| 1 | #pragma once |
| 2 | |
| 3 | #include <city.h> |
| 4 | #include <farmhash.h> |
| 5 | #include <metrohash.h> |
| 6 | #include <murmurhash2.h> |
| 7 | #include <murmurhash3.h> |
| 8 | |
| 9 | #include <Common/SipHash.h> |
| 10 | #include <Common/typeid_cast.h> |
| 11 | #include <Common/HashTable/Hash.h> |
| 12 | |
| 13 | #include "config_functions.h" |
| 14 | #if USE_XXHASH |
| 15 | # include <xxhash.h> |
| 16 | #endif |
| 17 | |
| 18 | #include "config_core.h" |
| 19 | #if USE_SSL |
| 20 | # include <openssl/md5.h> |
| 21 | # include <openssl/sha.h> |
| 22 | #endif |
| 23 | |
| 24 | #include <Poco/ByteOrder.h> |
| 25 | #include <DataTypes/DataTypesNumber.h> |
| 26 | #include <DataTypes/DataTypeString.h> |
| 27 | #include <DataTypes/DataTypeDate.h> |
| 28 | #include <DataTypes/DataTypeDateTime.h> |
| 29 | #include <DataTypes/DataTypeArray.h> |
| 30 | #include <DataTypes/DataTypeFixedString.h> |
| 31 | #include <DataTypes/DataTypeEnum.h> |
| 32 | #include <DataTypes/DataTypeTuple.h> |
| 33 | #include <Columns/ColumnsNumber.h> |
| 34 | #include <Columns/ColumnString.h> |
| 35 | #include <Columns/ColumnConst.h> |
| 36 | #include <Columns/ColumnFixedString.h> |
| 37 | #include <Columns/ColumnArray.h> |
| 38 | #include <Columns/ColumnTuple.h> |
| 39 | #include <Functions/IFunctionImpl.h> |
| 40 | #include <Functions/FunctionHelpers.h> |
| 41 | #include <ext/range.h> |
| 42 | #include <ext/bit_cast.h> |
| 43 | |
| 44 | |
| 45 | namespace DB |
| 46 | { |
| 47 | |
| 48 | namespace ErrorCodes |
| 49 | { |
| 50 | extern const int LOGICAL_ERROR; |
| 51 | extern const int NUMBER_OF_ARGUMENTS_DOESNT_MATCH; |
| 52 | extern const int NOT_IMPLEMENTED; |
| 53 | extern const int ILLEGAL_COLUMN; |
| 54 | } |
| 55 | |
| 56 | |
| 57 | /** Hashing functions. |
| 58 | * |
| 59 | * halfMD5: String -> UInt64 |
| 60 | * |
| 61 | * A faster cryptographic hash function: |
| 62 | * sipHash64: String -> UInt64 |
| 63 | * |
| 64 | * Fast non-cryptographic hash function for strings: |
| 65 | * cityHash64: String -> UInt64 |
| 66 | * |
| 67 | * A non-cryptographic hashes from a tuple of values of any types (uses respective function for strings and intHash64 for numbers): |
| 68 | * cityHash64: any* -> UInt64 |
| 69 | * sipHash64: any* -> UInt64 |
| 70 | * halfMD5: any* -> UInt64 |
| 71 | * |
| 72 | * Fast non-cryptographic hash function from any integer: |
| 73 | * intHash32: number -> UInt32 |
| 74 | * intHash64: number -> UInt64 |
| 75 | * |
| 76 | */ |
| 77 | |
| 78 | struct IntHash32Impl |
| 79 | { |
| 80 | using ReturnType = UInt32; |
| 81 | |
| 82 | static UInt32 apply(UInt64 x) |
| 83 | { |
| 84 | /// seed is taken from /dev/urandom. It allows you to avoid undesirable dependencies with hashes in different data structures. |
| 85 | return intHash32<0x75D9543DE018BF45ULL>(x); |
| 86 | } |
| 87 | }; |
| 88 | |
| 89 | struct IntHash64Impl |
| 90 | { |
| 91 | using ReturnType = UInt64; |
| 92 | |
| 93 | static UInt64 apply(UInt64 x) |
| 94 | { |
| 95 | return intHash64(x ^ 0x4CF2D2BAAE6DA887ULL); |
| 96 | } |
| 97 | }; |
| 98 | |
| 99 | #if USE_SSL |
| 100 | struct HalfMD5Impl |
| 101 | { |
| 102 | static constexpr auto name = "halfMD5" ; |
| 103 | using ReturnType = UInt64; |
| 104 | |
| 105 | static UInt64 apply(const char * begin, size_t size) |
| 106 | { |
| 107 | union |
| 108 | { |
| 109 | unsigned char char_data[16]; |
| 110 | uint64_t uint64_data; |
| 111 | } buf; |
| 112 | |
| 113 | MD5_CTX ctx; |
| 114 | MD5_Init(&ctx); |
| 115 | MD5_Update(&ctx, reinterpret_cast<const unsigned char *>(begin), size); |
| 116 | MD5_Final(buf.char_data, &ctx); |
| 117 | |
| 118 | return Poco::ByteOrder::flipBytes(static_cast<Poco::UInt64>(buf.uint64_data)); /// Compatibility with existing code. Cast need for old poco AND macos where UInt64 != uint64_t |
| 119 | } |
| 120 | |
| 121 | static UInt64 combineHashes(UInt64 h1, UInt64 h2) |
| 122 | { |
| 123 | UInt64 hashes[] = {h1, h2}; |
| 124 | return apply(reinterpret_cast<const char *>(hashes), 16); |
| 125 | } |
| 126 | |
| 127 | /// If true, it will use intHash32 or intHash64 to hash POD types. This behaviour is intended for better performance of some functions. |
| 128 | /// Otherwise it will hash bytes in memory as a string using corresponding hash function. |
| 129 | |
| 130 | static constexpr bool use_int_hash_for_pods = false; |
| 131 | }; |
| 132 | |
| 133 | struct MD5Impl |
| 134 | { |
| 135 | static constexpr auto name = "MD5" ; |
| 136 | enum { length = 16 }; |
| 137 | |
| 138 | static void apply(const char * begin, const size_t size, unsigned char * out_char_data) |
| 139 | { |
| 140 | MD5_CTX ctx; |
| 141 | MD5_Init(&ctx); |
| 142 | MD5_Update(&ctx, reinterpret_cast<const unsigned char *>(begin), size); |
| 143 | MD5_Final(out_char_data, &ctx); |
| 144 | } |
| 145 | }; |
| 146 | |
| 147 | struct SHA1Impl |
| 148 | { |
| 149 | static constexpr auto name = "SHA1" ; |
| 150 | enum { length = 20 }; |
| 151 | |
| 152 | static void apply(const char * begin, const size_t size, unsigned char * out_char_data) |
| 153 | { |
| 154 | SHA_CTX ctx; |
| 155 | SHA1_Init(&ctx); |
| 156 | SHA1_Update(&ctx, reinterpret_cast<const unsigned char *>(begin), size); |
| 157 | SHA1_Final(out_char_data, &ctx); |
| 158 | } |
| 159 | }; |
| 160 | |
| 161 | struct SHA224Impl |
| 162 | { |
| 163 | static constexpr auto name = "SHA224" ; |
| 164 | enum { length = 28 }; |
| 165 | |
| 166 | static void apply(const char * begin, const size_t size, unsigned char * out_char_data) |
| 167 | { |
| 168 | SHA256_CTX ctx; |
| 169 | SHA224_Init(&ctx); |
| 170 | SHA224_Update(&ctx, reinterpret_cast<const unsigned char *>(begin), size); |
| 171 | SHA224_Final(out_char_data, &ctx); |
| 172 | } |
| 173 | }; |
| 174 | |
| 175 | struct SHA256Impl |
| 176 | { |
| 177 | static constexpr auto name = "SHA256" ; |
| 178 | enum { length = 32 }; |
| 179 | |
| 180 | static void apply(const char * begin, const size_t size, unsigned char * out_char_data) |
| 181 | { |
| 182 | SHA256_CTX ctx; |
| 183 | SHA256_Init(&ctx); |
| 184 | SHA256_Update(&ctx, reinterpret_cast<const unsigned char *>(begin), size); |
| 185 | SHA256_Final(out_char_data, &ctx); |
| 186 | } |
| 187 | }; |
| 188 | #endif |
| 189 | |
| 190 | struct SipHash64Impl |
| 191 | { |
| 192 | static constexpr auto name = "sipHash64" ; |
| 193 | using ReturnType = UInt64; |
| 194 | |
| 195 | static UInt64 apply(const char * begin, size_t size) |
| 196 | { |
| 197 | return sipHash64(begin, size); |
| 198 | } |
| 199 | |
| 200 | static UInt64 combineHashes(UInt64 h1, UInt64 h2) |
| 201 | { |
| 202 | UInt64 hashes[] = {h1, h2}; |
| 203 | return apply(reinterpret_cast<const char *>(hashes), 16); |
| 204 | } |
| 205 | |
| 206 | static constexpr bool use_int_hash_for_pods = false; |
| 207 | }; |
| 208 | |
| 209 | struct SipHash128Impl |
| 210 | { |
| 211 | static constexpr auto name = "sipHash128" ; |
| 212 | enum { length = 16 }; |
| 213 | |
| 214 | static void apply(const char * begin, const size_t size, unsigned char * out_char_data) |
| 215 | { |
| 216 | sipHash128(begin, size, reinterpret_cast<char*>(out_char_data)); |
| 217 | } |
| 218 | }; |
| 219 | |
| 220 | |
| 221 | /** Why we need MurmurHash2? |
| 222 | * MurmurHash2 is an outdated hash function, superseded by MurmurHash3 and subsequently by CityHash, xxHash, HighwayHash. |
| 223 | * Usually there is no reason to use MurmurHash. |
| 224 | * It is needed for the cases when you already have MurmurHash in some applications and you want to reproduce it |
| 225 | * in ClickHouse as is. For example, it is needed to reproduce the behaviour |
| 226 | * for NGINX a/b testing module: https://nginx.ru/en/docs/http/ngx_http_split_clients_module.html |
| 227 | */ |
| 228 | struct MurmurHash2Impl32 |
| 229 | { |
| 230 | static constexpr auto name = "murmurHash2_32" ; |
| 231 | |
| 232 | using ReturnType = UInt32; |
| 233 | |
| 234 | static UInt32 apply(const char * data, const size_t size) |
| 235 | { |
| 236 | return MurmurHash2(data, size, 0); |
| 237 | } |
| 238 | |
| 239 | static UInt32 combineHashes(UInt32 h1, UInt32 h2) |
| 240 | { |
| 241 | return IntHash32Impl::apply(h1) ^ h2; |
| 242 | } |
| 243 | |
| 244 | static constexpr bool use_int_hash_for_pods = false; |
| 245 | }; |
| 246 | |
| 247 | struct MurmurHash2Impl64 |
| 248 | { |
| 249 | static constexpr auto name = "murmurHash2_64" ; |
| 250 | using ReturnType = UInt64; |
| 251 | |
| 252 | static UInt64 apply(const char * data, const size_t size) |
| 253 | { |
| 254 | return MurmurHash64A(data, size, 0); |
| 255 | } |
| 256 | |
| 257 | static UInt64 combineHashes(UInt64 h1, UInt64 h2) |
| 258 | { |
| 259 | return IntHash64Impl::apply(h1) ^ h2; |
| 260 | } |
| 261 | |
| 262 | static constexpr bool use_int_hash_for_pods = false; |
| 263 | }; |
| 264 | |
| 265 | /// To be compatible with gcc: https://github.com/gcc-mirror/gcc/blob/41d6b10e96a1de98e90a7c0378437c3255814b16/libstdc%2B%2B-v3/include/bits/functional_hash.h#L191 |
| 266 | struct GccMurmurHashImpl |
| 267 | { |
| 268 | static constexpr auto name = "gccMurmurHash" ; |
| 269 | using ReturnType = UInt64; |
| 270 | |
| 271 | static UInt64 apply(const char * data, const size_t size) |
| 272 | { |
| 273 | return MurmurHash64A(data, size, 0xc70f6907UL); |
| 274 | } |
| 275 | |
| 276 | static UInt64 combineHashes(UInt64 h1, UInt64 h2) |
| 277 | { |
| 278 | return IntHash64Impl::apply(h1) ^ h2; |
| 279 | } |
| 280 | |
| 281 | static constexpr bool use_int_hash_for_pods = false; |
| 282 | }; |
| 283 | |
| 284 | struct MurmurHash3Impl32 |
| 285 | { |
| 286 | static constexpr auto name = "murmurHash3_32" ; |
| 287 | using ReturnType = UInt32; |
| 288 | |
| 289 | static UInt32 apply(const char * data, const size_t size) |
| 290 | { |
| 291 | union |
| 292 | { |
| 293 | UInt32 h; |
| 294 | char bytes[sizeof(h)]; |
| 295 | }; |
| 296 | MurmurHash3_x86_32(data, size, 0, bytes); |
| 297 | return h; |
| 298 | } |
| 299 | |
| 300 | static UInt32 combineHashes(UInt32 h1, UInt32 h2) |
| 301 | { |
| 302 | return IntHash32Impl::apply(h1) ^ h2; |
| 303 | } |
| 304 | |
| 305 | static constexpr bool use_int_hash_for_pods = false; |
| 306 | }; |
| 307 | |
| 308 | struct MurmurHash3Impl64 |
| 309 | { |
| 310 | static constexpr auto name = "murmurHash3_64" ; |
| 311 | using ReturnType = UInt64; |
| 312 | |
| 313 | static UInt64 apply(const char * data, const size_t size) |
| 314 | { |
| 315 | union |
| 316 | { |
| 317 | UInt64 h[2]; |
| 318 | char bytes[16]; |
| 319 | }; |
| 320 | MurmurHash3_x64_128(data, size, 0, bytes); |
| 321 | return h[0] ^ h[1]; |
| 322 | } |
| 323 | |
| 324 | static UInt64 combineHashes(UInt64 h1, UInt64 h2) |
| 325 | { |
| 326 | return IntHash64Impl::apply(h1) ^ h2; |
| 327 | } |
| 328 | |
| 329 | static constexpr bool use_int_hash_for_pods = false; |
| 330 | }; |
| 331 | |
| 332 | /// http://hg.openjdk.java.net/jdk8u/jdk8u/jdk/file/478a4add975b/src/share/classes/java/lang/String.java#l1452 |
| 333 | /// Care should be taken to do all calculation in unsigned integers (to avoid undefined behaviour on overflow) |
| 334 | /// but obtain the same result as it is done in singed integers with two's complement arithmetic. |
| 335 | struct JavaHashImpl |
| 336 | { |
| 337 | static constexpr auto name = "javaHash" ; |
| 338 | using ReturnType = Int32; |
| 339 | |
| 340 | static Int32 apply(const char * data, const size_t size) |
| 341 | { |
| 342 | UInt32 h = 0; |
| 343 | for (size_t i = 0; i < size; ++i) |
| 344 | h = 31 * h + static_cast<UInt32>(static_cast<Int8>(data[i])); |
| 345 | return static_cast<Int32>(h); |
| 346 | } |
| 347 | |
| 348 | static Int32 combineHashes(Int32, Int32) |
| 349 | { |
| 350 | throw Exception("Java hash is not combineable for multiple arguments" , ErrorCodes::NOT_IMPLEMENTED); |
| 351 | } |
| 352 | |
| 353 | static constexpr bool use_int_hash_for_pods = false; |
| 354 | }; |
| 355 | |
| 356 | struct JavaHashUTF16LEImpl |
| 357 | { |
| 358 | static constexpr auto name = "javaHashUTF16LE" ; |
| 359 | using ReturnType = Int32; |
| 360 | |
| 361 | static Int32 apply(const char * raw_data, const size_t raw_size) |
| 362 | { |
| 363 | char * data = const_cast<char *>(raw_data); |
| 364 | size_t size = raw_size; |
| 365 | |
| 366 | // Remove Byte-order-mark(0xFFFE) for UTF-16LE |
| 367 | if (size >= 2 && data[0] == '\xFF' && data[1] == '\xFE') |
| 368 | { |
| 369 | data += 2; |
| 370 | size -= 2; |
| 371 | } |
| 372 | |
| 373 | if (size % 2 != 0) |
| 374 | throw Exception("Arguments for javaHashUTF16LE must be in the form of UTF-16" , ErrorCodes::LOGICAL_ERROR); |
| 375 | |
| 376 | UInt32 h = 0; |
| 377 | for (size_t i = 0; i < size; i += 2) |
| 378 | h = 31 * h + static_cast<UInt16>(static_cast<UInt8>(data[i]) | static_cast<UInt8>(data[i + 1]) << 8); |
| 379 | |
| 380 | return static_cast<Int32>(h); |
| 381 | } |
| 382 | |
| 383 | static Int32 combineHashes(Int32, Int32) |
| 384 | { |
| 385 | throw Exception("Java hash is not combineable for multiple arguments" , ErrorCodes::NOT_IMPLEMENTED); |
| 386 | } |
| 387 | |
| 388 | static constexpr bool use_int_hash_for_pods = false; |
| 389 | }; |
| 390 | |
| 391 | /// This is just JavaHash with zeroed out sign bit. |
| 392 | /// This function is used in Hive for versions before 3.0, |
| 393 | /// after 3.0, Hive uses murmur-hash3. |
| 394 | struct HiveHashImpl |
| 395 | { |
| 396 | static constexpr auto name = "hiveHash" ; |
| 397 | using ReturnType = Int32; |
| 398 | |
| 399 | static Int32 apply(const char * data, const size_t size) |
| 400 | { |
| 401 | return static_cast<Int32>(0x7FFFFFFF & static_cast<UInt32>(JavaHashImpl::apply(data, size))); |
| 402 | } |
| 403 | |
| 404 | static Int32 combineHashes(Int32, Int32) |
| 405 | { |
| 406 | throw Exception("Hive hash is not combineable for multiple arguments" , ErrorCodes::NOT_IMPLEMENTED); |
| 407 | } |
| 408 | |
| 409 | static constexpr bool use_int_hash_for_pods = false; |
| 410 | }; |
| 411 | |
| 412 | struct MurmurHash3Impl128 |
| 413 | { |
| 414 | static constexpr auto name = "murmurHash3_128" ; |
| 415 | enum { length = 16 }; |
| 416 | |
| 417 | static void apply(const char * begin, const size_t size, unsigned char * out_char_data) |
| 418 | { |
| 419 | MurmurHash3_x64_128(begin, size, 0, out_char_data); |
| 420 | } |
| 421 | }; |
| 422 | |
| 423 | struct ImplCityHash64 |
| 424 | { |
| 425 | static constexpr auto name = "cityHash64" ; |
| 426 | using ReturnType = UInt64; |
| 427 | using uint128_t = CityHash_v1_0_2::uint128; |
| 428 | |
| 429 | static auto combineHashes(UInt64 h1, UInt64 h2) { return CityHash_v1_0_2::Hash128to64(uint128_t(h1, h2)); } |
| 430 | static auto apply(const char * s, const size_t len) { return CityHash_v1_0_2::CityHash64(s, len); } |
| 431 | static constexpr bool use_int_hash_for_pods = true; |
| 432 | }; |
| 433 | |
| 434 | // see farmhash.h for definition of NAMESPACE_FOR_HASH_FUNCTIONS |
| 435 | struct ImplFarmHash64 |
| 436 | { |
| 437 | static constexpr auto name = "farmHash64" ; |
| 438 | using ReturnType = UInt64; |
| 439 | using uint128_t = NAMESPACE_FOR_HASH_FUNCTIONS::uint128_t; |
| 440 | |
| 441 | static auto combineHashes(UInt64 h1, UInt64 h2) { return NAMESPACE_FOR_HASH_FUNCTIONS::Hash128to64(uint128_t(h1, h2)); } |
| 442 | static auto apply(const char * s, const size_t len) { return NAMESPACE_FOR_HASH_FUNCTIONS::Hash64(s, len); } |
| 443 | static constexpr bool use_int_hash_for_pods = true; |
| 444 | }; |
| 445 | |
| 446 | struct ImplMetroHash64 |
| 447 | { |
| 448 | static constexpr auto name = "metroHash64" ; |
| 449 | using ReturnType = UInt64; |
| 450 | using uint128_t = CityHash_v1_0_2::uint128; |
| 451 | |
| 452 | static auto combineHashes(UInt64 h1, UInt64 h2) { return CityHash_v1_0_2::Hash128to64(uint128_t(h1, h2)); } |
| 453 | static auto apply(const char * s, const size_t len) |
| 454 | { |
| 455 | union |
| 456 | { |
| 457 | UInt64 u64; |
| 458 | UInt8 u8[sizeof(u64)]; |
| 459 | }; |
| 460 | |
| 461 | metrohash64_1(reinterpret_cast<const UInt8 *>(s), len, 0, u8); |
| 462 | |
| 463 | return u64; |
| 464 | } |
| 465 | |
| 466 | static constexpr bool use_int_hash_for_pods = true; |
| 467 | }; |
| 468 | |
| 469 | |
| 470 | #if USE_XXHASH |
| 471 | |
| 472 | struct ImplXxHash32 |
| 473 | { |
| 474 | static constexpr auto name = "xxHash32" ; |
| 475 | using ReturnType = UInt32; |
| 476 | |
| 477 | static auto apply(const char * s, const size_t len) { return XXH32(s, len, 0); } |
| 478 | /** |
| 479 | * With current implementation with more than 1 arguments it will give the results |
| 480 | * non-reproducable from outside of CH. |
| 481 | * |
| 482 | * Proper way of combining several input is to use streaming mode of hash function |
| 483 | * https://github.com/Cyan4973/xxHash/issues/114#issuecomment-334908566 |
| 484 | * |
| 485 | * In common case doable by init_state / update_state / finalize_state |
| 486 | */ |
| 487 | static auto combineHashes(UInt32 h1, UInt32 h2) { return IntHash32Impl::apply(h1) ^ h2; } |
| 488 | |
| 489 | static constexpr bool use_int_hash_for_pods = false; |
| 490 | }; |
| 491 | |
| 492 | |
| 493 | struct ImplXxHash64 |
| 494 | { |
| 495 | static constexpr auto name = "xxHash64" ; |
| 496 | using ReturnType = UInt64; |
| 497 | using uint128_t = CityHash_v1_0_2::uint128; |
| 498 | |
| 499 | static auto apply(const char * s, const size_t len) { return XXH64(s, len, 0); } |
| 500 | |
| 501 | /* |
| 502 | With current implementation with more than 1 arguments it will give the results |
| 503 | non-reproducable from outside of CH. (see comment on ImplXxHash32). |
| 504 | */ |
| 505 | static auto combineHashes(UInt64 h1, UInt64 h2) { return CityHash_v1_0_2::Hash128to64(uint128_t(h1, h2)); } |
| 506 | |
| 507 | static constexpr bool use_int_hash_for_pods = false; |
| 508 | }; |
| 509 | |
| 510 | #endif |
| 511 | |
| 512 | |
| 513 | template <typename Impl> |
| 514 | class FunctionStringHashFixedString : public IFunction |
| 515 | { |
| 516 | public: |
| 517 | static constexpr auto name = Impl::name; |
| 518 | static FunctionPtr create(const Context &) { return std::make_shared<FunctionStringHashFixedString>(); } |
| 519 | |
| 520 | String getName() const override |
| 521 | { |
| 522 | return name; |
| 523 | } |
| 524 | |
| 525 | size_t getNumberOfArguments() const override { return 1; } |
| 526 | |
| 527 | DataTypePtr getReturnTypeImpl(const DataTypes & arguments) const override |
| 528 | { |
| 529 | if (!isString(arguments[0])) |
| 530 | throw Exception("Illegal type " + arguments[0]->getName() + " of argument of function " + getName(), |
| 531 | ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT); |
| 532 | |
| 533 | return std::make_shared<DataTypeFixedString>(Impl::length); |
| 534 | } |
| 535 | |
| 536 | bool useDefaultImplementationForConstants() const override { return true; } |
| 537 | |
| 538 | void executeImpl(Block & block, const ColumnNumbers & arguments, size_t result, size_t /*input_rows_count*/) override |
| 539 | { |
| 540 | if (const ColumnString * col_from = checkAndGetColumn<ColumnString>(block.getByPosition(arguments[0]).column.get())) |
| 541 | { |
| 542 | auto col_to = ColumnFixedString::create(Impl::length); |
| 543 | |
| 544 | const typename ColumnString::Chars & data = col_from->getChars(); |
| 545 | const typename ColumnString::Offsets & offsets = col_from->getOffsets(); |
| 546 | auto & chars_to = col_to->getChars(); |
| 547 | const auto size = offsets.size(); |
| 548 | chars_to.resize(size * Impl::length); |
| 549 | |
| 550 | ColumnString::Offset current_offset = 0; |
| 551 | for (size_t i = 0; i < size; ++i) |
| 552 | { |
| 553 | Impl::apply( |
| 554 | reinterpret_cast<const char *>(&data[current_offset]), |
| 555 | offsets[i] - current_offset - 1, |
| 556 | &chars_to[i * Impl::length]); |
| 557 | |
| 558 | current_offset = offsets[i]; |
| 559 | } |
| 560 | |
| 561 | block.getByPosition(result).column = std::move(col_to); |
| 562 | } |
| 563 | else |
| 564 | throw Exception("Illegal column " + block.getByPosition(arguments[0]).column->getName() |
| 565 | + " of first argument of function " + getName(), |
| 566 | ErrorCodes::ILLEGAL_COLUMN); |
| 567 | } |
| 568 | }; |
| 569 | |
| 570 | |
| 571 | template <typename Impl, typename Name> |
| 572 | class FunctionIntHash : public IFunction |
| 573 | { |
| 574 | public: |
| 575 | static constexpr auto name = Name::name; |
| 576 | static FunctionPtr create(const Context &) { return std::make_shared<FunctionIntHash>(); } |
| 577 | |
| 578 | private: |
| 579 | using ToType = typename Impl::ReturnType; |
| 580 | |
| 581 | template <typename FromType> |
| 582 | void executeType(Block & block, const ColumnNumbers & arguments, size_t result) |
| 583 | { |
| 584 | if (auto col_from = checkAndGetColumn<ColumnVector<FromType>>(block.getByPosition(arguments[0]).column.get())) |
| 585 | { |
| 586 | auto col_to = ColumnVector<ToType>::create(); |
| 587 | |
| 588 | const typename ColumnVector<FromType>::Container & vec_from = col_from->getData(); |
| 589 | typename ColumnVector<ToType>::Container & vec_to = col_to->getData(); |
| 590 | |
| 591 | size_t size = vec_from.size(); |
| 592 | vec_to.resize(size); |
| 593 | for (size_t i = 0; i < size; ++i) |
| 594 | vec_to[i] = Impl::apply(vec_from[i]); |
| 595 | |
| 596 | block.getByPosition(result).column = std::move(col_to); |
| 597 | } |
| 598 | else |
| 599 | throw Exception("Illegal column " + block.getByPosition(arguments[0]).column->getName() |
| 600 | + " of first argument of function " + Name::name, |
| 601 | ErrorCodes::ILLEGAL_COLUMN); |
| 602 | } |
| 603 | |
| 604 | public: |
| 605 | String getName() const override |
| 606 | { |
| 607 | return name; |
| 608 | } |
| 609 | |
| 610 | size_t getNumberOfArguments() const override { return 1; } |
| 611 | |
| 612 | DataTypePtr getReturnTypeImpl(const DataTypes & arguments) const override |
| 613 | { |
| 614 | if (!arguments[0]->isValueRepresentedByNumber()) |
| 615 | throw Exception("Illegal type " + arguments[0]->getName() + " of argument of function " + getName(), |
| 616 | ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT); |
| 617 | |
| 618 | return std::make_shared<DataTypeNumber<typename Impl::ReturnType>>(); |
| 619 | } |
| 620 | |
| 621 | bool useDefaultImplementationForConstants() const override { return true; } |
| 622 | |
| 623 | void executeImpl(Block & block, const ColumnNumbers & arguments, size_t result, size_t /*input_rows_count*/) override |
| 624 | { |
| 625 | const IDataType * from_type = block.getByPosition(arguments[0]).type.get(); |
| 626 | WhichDataType which(from_type); |
| 627 | |
| 628 | if (which.isUInt8()) executeType<UInt8>(block, arguments, result); |
| 629 | else if (which.isUInt16()) executeType<UInt16>(block, arguments, result); |
| 630 | else if (which.isUInt32()) executeType<UInt32>(block, arguments, result); |
| 631 | else if (which.isUInt64()) executeType<UInt64>(block, arguments, result); |
| 632 | else if (which.isInt8()) executeType<Int8>(block, arguments, result); |
| 633 | else if (which.isInt16()) executeType<Int16>(block, arguments, result); |
| 634 | else if (which.isInt32()) executeType<Int32>(block, arguments, result); |
| 635 | else if (which.isInt64()) executeType<Int64>(block, arguments, result); |
| 636 | else if (which.isDate()) executeType<UInt16>(block, arguments, result); |
| 637 | else if (which.isDateTime()) executeType<UInt32>(block, arguments, result); |
| 638 | else |
| 639 | throw Exception("Illegal type " + block.getByPosition(arguments[0]).type->getName() + " of argument of function " + getName(), |
| 640 | ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT); |
| 641 | } |
| 642 | }; |
| 643 | |
| 644 | |
| 645 | template <typename Impl> |
| 646 | class FunctionAnyHash : public IFunction |
| 647 | { |
| 648 | public: |
| 649 | static constexpr auto name = Impl::name; |
| 650 | static FunctionPtr create(const Context &) { return std::make_shared<FunctionAnyHash>(); } |
| 651 | |
| 652 | private: |
| 653 | using ToType = typename Impl::ReturnType; |
| 654 | |
| 655 | template <typename FromType, bool first> |
| 656 | void executeIntType(const IColumn * column, typename ColumnVector<ToType>::Container & vec_to) |
| 657 | { |
| 658 | if (const ColumnVector<FromType> * col_from = checkAndGetColumn<ColumnVector<FromType>>(column)) |
| 659 | { |
| 660 | const typename ColumnVector<FromType>::Container & vec_from = col_from->getData(); |
| 661 | size_t size = vec_from.size(); |
| 662 | for (size_t i = 0; i < size; ++i) |
| 663 | { |
| 664 | ToType h; |
| 665 | |
| 666 | if constexpr (Impl::use_int_hash_for_pods) |
| 667 | { |
| 668 | if constexpr (std::is_same_v<ToType, UInt64>) |
| 669 | h = IntHash64Impl::apply(ext::bit_cast<UInt64>(vec_from[i])); |
| 670 | else |
| 671 | h = IntHash32Impl::apply(ext::bit_cast<UInt32>(vec_from[i])); |
| 672 | } |
| 673 | else |
| 674 | { |
| 675 | h = Impl::apply(reinterpret_cast<const char *>(&vec_from[i]), sizeof(vec_from[i])); |
| 676 | } |
| 677 | |
| 678 | if (first) |
| 679 | vec_to[i] = h; |
| 680 | else |
| 681 | vec_to[i] = Impl::combineHashes(vec_to[i], h); |
| 682 | } |
| 683 | } |
| 684 | else if (auto col_from_const = checkAndGetColumnConst<ColumnVector<FromType>>(column)) |
| 685 | { |
| 686 | auto value = col_from_const->template getValue<FromType>(); |
| 687 | ToType hash; |
| 688 | if constexpr (std::is_same_v<ToType, UInt64>) |
| 689 | hash = IntHash64Impl::apply(ext::bit_cast<UInt64>(value)); |
| 690 | else |
| 691 | hash = IntHash32Impl::apply(ext::bit_cast<UInt32>(value)); |
| 692 | |
| 693 | size_t size = vec_to.size(); |
| 694 | if (first) |
| 695 | { |
| 696 | vec_to.assign(size, hash); |
| 697 | } |
| 698 | else |
| 699 | { |
| 700 | for (size_t i = 0; i < size; ++i) |
| 701 | vec_to[i] = Impl::combineHashes(vec_to[i], hash); |
| 702 | } |
| 703 | } |
| 704 | else |
| 705 | throw Exception("Illegal column " + column->getName() |
| 706 | + " of argument of function " + getName(), |
| 707 | ErrorCodes::ILLEGAL_COLUMN); |
| 708 | } |
| 709 | |
| 710 | template <bool first> |
| 711 | void executeGeneric(const IColumn * column, typename ColumnVector<ToType>::Container & vec_to) |
| 712 | { |
| 713 | for (size_t i = 0, size = column->size(); i < size; ++i) |
| 714 | { |
| 715 | StringRef bytes = column->getDataAt(i); |
| 716 | const ToType h = Impl::apply(bytes.data, bytes.size); |
| 717 | if (first) |
| 718 | vec_to[i] = h; |
| 719 | else |
| 720 | vec_to[i] = Impl::combineHashes(vec_to[i], h); |
| 721 | } |
| 722 | } |
| 723 | |
| 724 | template <bool first> |
| 725 | void executeString(const IColumn * column, typename ColumnVector<ToType>::Container & vec_to) |
| 726 | { |
| 727 | if (const ColumnString * col_from = checkAndGetColumn<ColumnString>(column)) |
| 728 | { |
| 729 | const typename ColumnString::Chars & data = col_from->getChars(); |
| 730 | const typename ColumnString::Offsets & offsets = col_from->getOffsets(); |
| 731 | size_t size = offsets.size(); |
| 732 | |
| 733 | ColumnString::Offset current_offset = 0; |
| 734 | for (size_t i = 0; i < size; ++i) |
| 735 | { |
| 736 | const ToType h = Impl::apply( |
| 737 | reinterpret_cast<const char *>(&data[current_offset]), |
| 738 | offsets[i] - current_offset - 1); |
| 739 | |
| 740 | if (first) |
| 741 | vec_to[i] = h; |
| 742 | else |
| 743 | vec_to[i] = Impl::combineHashes(vec_to[i], h); |
| 744 | |
| 745 | current_offset = offsets[i]; |
| 746 | } |
| 747 | } |
| 748 | else if (const ColumnFixedString * col_from_fixed = checkAndGetColumn<ColumnFixedString>(column)) |
| 749 | { |
| 750 | const typename ColumnString::Chars & data = col_from_fixed->getChars(); |
| 751 | size_t n = col_from_fixed->getN(); |
| 752 | size_t size = data.size() / n; |
| 753 | |
| 754 | for (size_t i = 0; i < size; ++i) |
| 755 | { |
| 756 | const ToType h = Impl::apply(reinterpret_cast<const char *>(&data[i * n]), n); |
| 757 | if (first) |
| 758 | vec_to[i] = h; |
| 759 | else |
| 760 | vec_to[i] = Impl::combineHashes(vec_to[i], h); |
| 761 | } |
| 762 | } |
| 763 | else if (const ColumnConst * col_from_const = checkAndGetColumnConstStringOrFixedString(column)) |
| 764 | { |
| 765 | String value = col_from_const->getValue<String>().data(); |
| 766 | const ToType hash = Impl::apply(value.data(), value.size()); |
| 767 | const size_t size = vec_to.size(); |
| 768 | |
| 769 | if (first) |
| 770 | { |
| 771 | vec_to.assign(size, hash); |
| 772 | } |
| 773 | else |
| 774 | { |
| 775 | for (size_t i = 0; i < size; ++i) |
| 776 | { |
| 777 | vec_to[i] = Impl::combineHashes(vec_to[i], hash); |
| 778 | } |
| 779 | } |
| 780 | } |
| 781 | else |
| 782 | throw Exception("Illegal column " + column->getName() |
| 783 | + " of first argument of function " + getName(), |
| 784 | ErrorCodes::ILLEGAL_COLUMN); |
| 785 | } |
| 786 | |
| 787 | template <bool first> |
| 788 | void executeArray(const IDataType * type, const IColumn * column, typename ColumnVector<ToType>::Container & vec_to) |
| 789 | { |
| 790 | const IDataType * nested_type = typeid_cast<const DataTypeArray *>(type)->getNestedType().get(); |
| 791 | |
| 792 | if (const ColumnArray * col_from = checkAndGetColumn<ColumnArray>(column)) |
| 793 | { |
| 794 | const IColumn * nested_column = &col_from->getData(); |
| 795 | const ColumnArray::Offsets & offsets = col_from->getOffsets(); |
| 796 | const size_t nested_size = nested_column->size(); |
| 797 | |
| 798 | typename ColumnVector<ToType>::Container vec_temp(nested_size); |
| 799 | executeAny<true>(nested_type, nested_column, vec_temp); |
| 800 | |
| 801 | const size_t size = offsets.size(); |
| 802 | |
| 803 | ColumnArray::Offset current_offset = 0; |
| 804 | for (size_t i = 0; i < size; ++i) |
| 805 | { |
| 806 | ColumnArray::Offset next_offset = offsets[i]; |
| 807 | |
| 808 | ToType h; |
| 809 | if constexpr (std::is_same_v<ToType, UInt64>) |
| 810 | h = IntHash64Impl::apply(next_offset - current_offset); |
| 811 | else |
| 812 | h = IntHash32Impl::apply(next_offset - current_offset); |
| 813 | |
| 814 | if (first) |
| 815 | vec_to[i] = h; |
| 816 | else |
| 817 | vec_to[i] = Impl::combineHashes(vec_to[i], h); |
| 818 | |
| 819 | for (size_t j = current_offset; j < next_offset; ++j) |
| 820 | vec_to[i] = Impl::combineHashes(vec_to[i], vec_temp[j]); |
| 821 | |
| 822 | current_offset = offsets[i]; |
| 823 | } |
| 824 | } |
| 825 | else if (const ColumnConst * col_from_const = checkAndGetColumnConst<ColumnArray>(column)) |
| 826 | { |
| 827 | /// NOTE: here, of course, you can do without the materialization of the column. |
| 828 | ColumnPtr full_column = col_from_const->convertToFullColumn(); |
| 829 | executeArray<first>(type, &*full_column, vec_to); |
| 830 | } |
| 831 | else |
| 832 | throw Exception("Illegal column " + column->getName() |
| 833 | + " of first argument of function " + getName(), |
| 834 | ErrorCodes::ILLEGAL_COLUMN); |
| 835 | } |
| 836 | |
| 837 | template <bool first> |
| 838 | void executeAny(const IDataType * from_type, const IColumn * icolumn, typename ColumnVector<ToType>::Container & vec_to) |
| 839 | { |
| 840 | WhichDataType which(from_type); |
| 841 | |
| 842 | if (which.isUInt8()) executeIntType<UInt8, first>(icolumn, vec_to); |
| 843 | else if (which.isUInt16()) executeIntType<UInt16, first>(icolumn, vec_to); |
| 844 | else if (which.isUInt32()) executeIntType<UInt32, first>(icolumn, vec_to); |
| 845 | else if (which.isUInt64()) executeIntType<UInt64, first>(icolumn, vec_to); |
| 846 | else if (which.isInt8()) executeIntType<Int8, first>(icolumn, vec_to); |
| 847 | else if (which.isInt16()) executeIntType<Int16, first>(icolumn, vec_to); |
| 848 | else if (which.isInt32()) executeIntType<Int32, first>(icolumn, vec_to); |
| 849 | else if (which.isInt64()) executeIntType<Int64, first>(icolumn, vec_to); |
| 850 | else if (which.isEnum8()) executeIntType<Int8, first>(icolumn, vec_to); |
| 851 | else if (which.isEnum16()) executeIntType<Int16, first>(icolumn, vec_to); |
| 852 | else if (which.isDate()) executeIntType<UInt16, first>(icolumn, vec_to); |
| 853 | else if (which.isDateTime()) executeIntType<UInt32, first>(icolumn, vec_to); |
| 854 | else if (which.isFloat32()) executeIntType<Float32, first>(icolumn, vec_to); |
| 855 | else if (which.isFloat64()) executeIntType<Float64, first>(icolumn, vec_to); |
| 856 | else if (which.isString()) executeString<first>(icolumn, vec_to); |
| 857 | else if (which.isFixedString()) executeString<first>(icolumn, vec_to); |
| 858 | else if (which.isArray()) executeArray<first>(from_type, icolumn, vec_to); |
| 859 | else |
| 860 | executeGeneric<first>(icolumn, vec_to); |
| 861 | } |
| 862 | |
| 863 | void executeForArgument(const IDataType * type, const IColumn * column, typename ColumnVector<ToType>::Container & vec_to, bool & is_first) |
| 864 | { |
| 865 | /// Flattening of tuples. |
| 866 | if (const ColumnTuple * tuple = typeid_cast<const ColumnTuple *>(column)) |
| 867 | { |
| 868 | const auto & tuple_columns = tuple->getColumns(); |
| 869 | const DataTypes & tuple_types = typeid_cast<const DataTypeTuple &>(*type).getElements(); |
| 870 | size_t tuple_size = tuple_columns.size(); |
| 871 | for (size_t i = 0; i < tuple_size; ++i) |
| 872 | executeForArgument(tuple_types[i].get(), tuple_columns[i].get(), vec_to, is_first); |
| 873 | } |
| 874 | else if (const ColumnTuple * tuple_const = checkAndGetColumnConstData<ColumnTuple>(column)) |
| 875 | { |
| 876 | const auto & tuple_columns = tuple_const->getColumns(); |
| 877 | const DataTypes & tuple_types = typeid_cast<const DataTypeTuple &>(*type).getElements(); |
| 878 | size_t tuple_size = tuple_columns.size(); |
| 879 | for (size_t i = 0; i < tuple_size; ++i) |
| 880 | { |
| 881 | auto tmp = ColumnConst::create(tuple_columns[i], column->size()); |
| 882 | executeForArgument(tuple_types[i].get(), tmp.get(), vec_to, is_first); |
| 883 | } |
| 884 | } |
| 885 | else |
| 886 | { |
| 887 | if (is_first) |
| 888 | executeAny<true>(type, column, vec_to); |
| 889 | else |
| 890 | executeAny<false>(type, column, vec_to); |
| 891 | } |
| 892 | |
| 893 | is_first = false; |
| 894 | } |
| 895 | |
| 896 | public: |
| 897 | String getName() const override |
| 898 | { |
| 899 | return name; |
| 900 | } |
| 901 | |
| 902 | bool isVariadic() const override { return true; } |
| 903 | size_t getNumberOfArguments() const override { return 0; } |
| 904 | bool useDefaultImplementationForConstants() const override { return true; } |
| 905 | |
| 906 | DataTypePtr getReturnTypeImpl(const DataTypes & /*arguments*/) const override |
| 907 | { |
| 908 | return std::make_shared<DataTypeNumber<ToType>>(); |
| 909 | } |
| 910 | |
| 911 | void executeImpl(Block & block, const ColumnNumbers & arguments, size_t result, size_t input_rows_count) override |
| 912 | { |
| 913 | size_t rows = input_rows_count; |
| 914 | auto col_to = ColumnVector<ToType>::create(rows); |
| 915 | |
| 916 | typename ColumnVector<ToType>::Container & vec_to = col_to->getData(); |
| 917 | |
| 918 | if (arguments.empty()) |
| 919 | { |
| 920 | /// Constant random number from /dev/urandom is used as a hash value of empty list of arguments. |
| 921 | vec_to.assign(rows, static_cast<ToType>(0xe28dbde7fe22e41c)); |
| 922 | } |
| 923 | |
| 924 | /// The function supports arbitrary number of arguments of arbitrary types. |
| 925 | |
| 926 | bool is_first_argument = true; |
| 927 | for (size_t i = 0; i < arguments.size(); ++i) |
| 928 | { |
| 929 | const ColumnWithTypeAndName & col = block.getByPosition(arguments[i]); |
| 930 | executeForArgument(col.type.get(), col.column.get(), vec_to, is_first_argument); |
| 931 | } |
| 932 | |
| 933 | block.getByPosition(result).column = std::move(col_to); |
| 934 | } |
| 935 | }; |
| 936 | |
| 937 | |
| 938 | struct URLHashImpl |
| 939 | { |
| 940 | static UInt64 apply(const char * data, const size_t size) |
| 941 | { |
| 942 | /// do not take last slash, '?' or '#' character into account |
| 943 | if (size > 0 && (data[size - 1] == '/' || data[size - 1] == '?' || data[size - 1] == '#')) |
| 944 | return CityHash_v1_0_2::CityHash64(data, size - 1); |
| 945 | |
| 946 | return CityHash_v1_0_2::CityHash64(data, size); |
| 947 | } |
| 948 | }; |
| 949 | |
| 950 | |
| 951 | struct URLHierarchyHashImpl |
| 952 | { |
| 953 | static size_t findLevelLength(const UInt64 level, const char * begin, const char * end) |
| 954 | { |
| 955 | auto pos = begin; |
| 956 | |
| 957 | /// Let's parse everything that goes before the path |
| 958 | |
| 959 | /// Suppose that the protocol has already been changed to lowercase. |
| 960 | while (pos < end && ((*pos > 'a' && *pos < 'z') || (*pos > '0' && *pos < '9'))) |
| 961 | ++pos; |
| 962 | |
| 963 | /** We will calculate the hierarchy only for URLs in which there is a protocol, and after it there are two slashes. |
| 964 | * (http, file - fit, mailto, magnet - do not fit), and after two slashes there is still something |
| 965 | * For the rest, simply return the full URL as the only element of the hierarchy. |
| 966 | */ |
| 967 | if (pos == begin || pos == end || !(*pos++ == ':' && pos < end && *pos++ == '/' && pos < end && *pos++ == '/' && pos < end)) |
| 968 | { |
| 969 | pos = end; |
| 970 | return 0 == level ? pos - begin : 0; |
| 971 | } |
| 972 | |
| 973 | /// The domain for simplicity is everything that after the protocol and the two slashes, until the next slash or before `?` or `#` |
| 974 | while (pos < end && !(*pos == '/' || *pos == '?' || *pos == '#')) |
| 975 | ++pos; |
| 976 | |
| 977 | if (pos != end) |
| 978 | ++pos; |
| 979 | |
| 980 | if (0 == level) |
| 981 | return pos - begin; |
| 982 | |
| 983 | UInt64 current_level = 0; |
| 984 | |
| 985 | while (current_level != level && pos < end) |
| 986 | { |
| 987 | /// We go to the next `/` or `?` or `#`, skipping all at the beginning. |
| 988 | while (pos < end && (*pos == '/' || *pos == '?' || *pos == '#')) |
| 989 | ++pos; |
| 990 | if (pos == end) |
| 991 | break; |
| 992 | while (pos < end && !(*pos == '/' || *pos == '?' || *pos == '#')) |
| 993 | ++pos; |
| 994 | |
| 995 | if (pos != end) |
| 996 | ++pos; |
| 997 | |
| 998 | ++current_level; |
| 999 | } |
| 1000 | |
| 1001 | return current_level == level ? pos - begin : 0; |
| 1002 | } |
| 1003 | |
| 1004 | static UInt64 apply(const UInt64 level, const char * data, const size_t size) |
| 1005 | { |
| 1006 | return URLHashImpl::apply(data, findLevelLength(level, data, data + size)); |
| 1007 | } |
| 1008 | }; |
| 1009 | |
| 1010 | |
| 1011 | class FunctionURLHash : public IFunction |
| 1012 | { |
| 1013 | public: |
| 1014 | static constexpr auto name = "URLHash" ; |
| 1015 | static FunctionPtr create(const Context &) { return std::make_shared<FunctionURLHash>(); } |
| 1016 | |
| 1017 | String getName() const override { return name; } |
| 1018 | |
| 1019 | bool isVariadic() const override { return true; } |
| 1020 | size_t getNumberOfArguments() const override { return 0; } |
| 1021 | |
| 1022 | DataTypePtr getReturnTypeImpl(const DataTypes & arguments) const override |
| 1023 | { |
| 1024 | const auto arg_count = arguments.size(); |
| 1025 | if (arg_count != 1 && arg_count != 2) |
| 1026 | throw Exception{"Number of arguments for function " + getName() + " doesn't match: passed " + |
| 1027 | toString(arg_count) + ", should be 1 or 2." , ErrorCodes::NUMBER_OF_ARGUMENTS_DOESNT_MATCH}; |
| 1028 | |
| 1029 | const auto first_arg = arguments.front().get(); |
| 1030 | if (!WhichDataType(first_arg).isString()) |
| 1031 | throw Exception{"Illegal type " + first_arg->getName() + " of argument of function " + getName(), ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT}; |
| 1032 | |
| 1033 | if (arg_count == 2) |
| 1034 | { |
| 1035 | const auto & second_arg = arguments.back(); |
| 1036 | if (!isInteger(second_arg)) |
| 1037 | throw Exception{"Illegal type " + second_arg->getName() + " of argument of function " + getName(), ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT}; |
| 1038 | } |
| 1039 | |
| 1040 | return std::make_shared<DataTypeUInt64>(); |
| 1041 | } |
| 1042 | |
| 1043 | bool useDefaultImplementationForConstants() const override { return true; } |
| 1044 | ColumnNumbers getArgumentsThatAreAlwaysConstant() const override { return {1}; } |
| 1045 | |
| 1046 | void executeImpl(Block & block, const ColumnNumbers & arguments, size_t result, size_t /*input_rows_count*/) override |
| 1047 | { |
| 1048 | const auto arg_count = arguments.size(); |
| 1049 | |
| 1050 | if (arg_count == 1) |
| 1051 | executeSingleArg(block, arguments, result); |
| 1052 | else if (arg_count == 2) |
| 1053 | executeTwoArgs(block, arguments, result); |
| 1054 | else |
| 1055 | throw Exception{"got into IFunction::execute with unexpected number of arguments" , ErrorCodes::LOGICAL_ERROR}; |
| 1056 | } |
| 1057 | |
| 1058 | private: |
| 1059 | void executeSingleArg(Block & block, const ColumnNumbers & arguments, const size_t result) const |
| 1060 | { |
| 1061 | const auto col_untyped = block.getByPosition(arguments.front()).column.get(); |
| 1062 | |
| 1063 | if (const auto col_from = checkAndGetColumn<ColumnString>(col_untyped)) |
| 1064 | { |
| 1065 | const auto size = col_from->size(); |
| 1066 | auto col_to = ColumnUInt64::create(size); |
| 1067 | |
| 1068 | const auto & chars = col_from->getChars(); |
| 1069 | const auto & offsets = col_from->getOffsets(); |
| 1070 | auto & out = col_to->getData(); |
| 1071 | |
| 1072 | ColumnString::Offset current_offset = 0; |
| 1073 | for (size_t i = 0; i < size; ++i) |
| 1074 | { |
| 1075 | out[i] = URLHashImpl::apply( |
| 1076 | reinterpret_cast<const char *>(&chars[current_offset]), |
| 1077 | offsets[i] - current_offset - 1); |
| 1078 | |
| 1079 | current_offset = offsets[i]; |
| 1080 | } |
| 1081 | |
| 1082 | block.getByPosition(result).column = std::move(col_to); |
| 1083 | } |
| 1084 | else |
| 1085 | throw Exception{"Illegal column " + block.getByPosition(arguments[0]).column->getName() + |
| 1086 | " of argument of function " + getName(), ErrorCodes::ILLEGAL_COLUMN}; |
| 1087 | } |
| 1088 | |
| 1089 | void executeTwoArgs(Block & block, const ColumnNumbers & arguments, const size_t result) const |
| 1090 | { |
| 1091 | const auto level_col = block.getByPosition(arguments.back()).column.get(); |
| 1092 | if (!isColumnConst(*level_col)) |
| 1093 | throw Exception{"Second argument of function " + getName() + " must be an integral constant" , ErrorCodes::ILLEGAL_COLUMN}; |
| 1094 | |
| 1095 | const auto level = level_col->get64(0); |
| 1096 | |
| 1097 | const auto col_untyped = block.getByPosition(arguments.front()).column.get(); |
| 1098 | if (const auto col_from = checkAndGetColumn<ColumnString>(col_untyped)) |
| 1099 | { |
| 1100 | const auto size = col_from->size(); |
| 1101 | auto col_to = ColumnUInt64::create(size); |
| 1102 | |
| 1103 | const auto & chars = col_from->getChars(); |
| 1104 | const auto & offsets = col_from->getOffsets(); |
| 1105 | auto & out = col_to->getData(); |
| 1106 | |
| 1107 | ColumnString::Offset current_offset = 0; |
| 1108 | for (size_t i = 0; i < size; ++i) |
| 1109 | { |
| 1110 | out[i] = URLHierarchyHashImpl::apply( |
| 1111 | level, |
| 1112 | reinterpret_cast<const char *>(&chars[current_offset]), |
| 1113 | offsets[i] - current_offset - 1); |
| 1114 | |
| 1115 | current_offset = offsets[i]; |
| 1116 | } |
| 1117 | |
| 1118 | block.getByPosition(result).column = std::move(col_to); |
| 1119 | } |
| 1120 | else |
| 1121 | throw Exception{"Illegal column " + block.getByPosition(arguments[0]).column->getName() + |
| 1122 | " of argument of function " + getName(), ErrorCodes::ILLEGAL_COLUMN}; |
| 1123 | } |
| 1124 | }; |
| 1125 | |
| 1126 | |
| 1127 | struct NameIntHash32 { static constexpr auto name = "intHash32" ; }; |
| 1128 | struct NameIntHash64 { static constexpr auto name = "intHash64" ; }; |
| 1129 | |
| 1130 | #if USE_SSL |
| 1131 | using FunctionHalfMD5 = FunctionAnyHash<HalfMD5Impl>; |
| 1132 | #endif |
| 1133 | using FunctionSipHash64 = FunctionAnyHash<SipHash64Impl>; |
| 1134 | using FunctionIntHash32 = FunctionIntHash<IntHash32Impl, NameIntHash32>; |
| 1135 | using FunctionIntHash64 = FunctionIntHash<IntHash64Impl, NameIntHash64>; |
| 1136 | #if USE_SSL |
| 1137 | using FunctionMD5 = FunctionStringHashFixedString<MD5Impl>; |
| 1138 | using FunctionSHA1 = FunctionStringHashFixedString<SHA1Impl>; |
| 1139 | using FunctionSHA224 = FunctionStringHashFixedString<SHA224Impl>; |
| 1140 | using FunctionSHA256 = FunctionStringHashFixedString<SHA256Impl>; |
| 1141 | #endif |
| 1142 | using FunctionSipHash128 = FunctionStringHashFixedString<SipHash128Impl>; |
| 1143 | using FunctionCityHash64 = FunctionAnyHash<ImplCityHash64>; |
| 1144 | using FunctionFarmHash64 = FunctionAnyHash<ImplFarmHash64>; |
| 1145 | using FunctionMetroHash64 = FunctionAnyHash<ImplMetroHash64>; |
| 1146 | using FunctionMurmurHash2_32 = FunctionAnyHash<MurmurHash2Impl32>; |
| 1147 | using FunctionMurmurHash2_64 = FunctionAnyHash<MurmurHash2Impl64>; |
| 1148 | using FunctionGccMurmurHash = FunctionAnyHash<GccMurmurHashImpl>; |
| 1149 | using FunctionMurmurHash3_32 = FunctionAnyHash<MurmurHash3Impl32>; |
| 1150 | using FunctionMurmurHash3_64 = FunctionAnyHash<MurmurHash3Impl64>; |
| 1151 | using FunctionMurmurHash3_128 = FunctionStringHashFixedString<MurmurHash3Impl128>; |
| 1152 | using FunctionJavaHash = FunctionAnyHash<JavaHashImpl>; |
| 1153 | using FunctionJavaHashUTF16LE = FunctionAnyHash<JavaHashUTF16LEImpl>; |
| 1154 | using FunctionHiveHash = FunctionAnyHash<HiveHashImpl>; |
| 1155 | |
| 1156 | #if USE_XXHASH |
| 1157 | using FunctionXxHash32 = FunctionAnyHash<ImplXxHash32>; |
| 1158 | using FunctionXxHash64 = FunctionAnyHash<ImplXxHash64>; |
| 1159 | #endif |
| 1160 | |
| 1161 | } |
| 1162 | |