| 1 |
|
| 2 |
# |
| 3 |
# InnoDB FULLTEXT SEARCH: CJK support - ngram parser |
| 4 |
# |
| 5 |
|
| 6 |
--source include/no_valgrind_without_big.inc |
| 7 |
--source include/have_simple_parser.inc |
| 8 |
|
| 9 |
SET NAMES utf8mb4; |
| 10 |
-- echo # Test Case 1: Test Chinese with GB2312 Charset. |
| 11 |
CREATE TABLE articles ( |
| 12 |
id INT UNSIGNED AUTO_INCREMENT NOT NULL PRIMARY KEY, |
| 13 |
title VARCHAR(200), |
| 14 |
body TEXT |
| 15 |
) ENGINE=InnoDB DEFAULT CHARACTER SET gb2312 COLLATE gb2312_chinese_ci; |
| 16 |
|
| 17 |
INSERT INTO articles (title, body) VALUES |
| 18 |
('数据库是数据的结构化集合','它可以是任何东西'), |
| 19 |
('从简单的购物清单到画展','或企业网络中的海量数据'); |
| 20 |
|
| 21 |
ALTER TABLE articles ADD FULLTEXT INDEX (title, body) WITH PARSER ngram; |
| 22 |
|
| 23 |
INSERT INTO articles (title, body) VALUES |
| 24 |
('要想将数据添加到数据库','或访问、处理计算机数据库中保存的数据'), |
| 25 |
('需要使用数据库管理系统','计算机是处理大量数据的理想工具'); |
| 26 |
|
| 27 |
--sorted_result |
| 28 |
SELECT * FROM articles WHERE |
| 29 |
MATCH(title, body) AGAINST('数据'); |
| 30 |
|
| 31 |
--sorted_result |
| 32 |
SELECT * FROM articles WHERE |
| 33 |
MATCH(title, body) AGAINST('数*' IN BOOLEAN MODE); |
| 34 |
|
| 35 |
SELECT * FROM articles WHERE |
| 36 |
MATCH(title, body) AGAINST('+数据库 -计算机' IN BOOLEAN MODE); |
| 37 |
|
| 38 |
SELECT * FROM articles WHERE |
| 39 |
MATCH(title, body) AGAINST('+(数据库 处理) -计算机' IN BOOLEAN MODE); |
| 40 |
|
| 41 |
--sorted_result |
| 42 |
SELECT * FROM articles WHERE |
| 43 |
MATCH(title, body) AGAINST('数据库'); |
| 44 |
|
| 45 |
--sorted_result |
| 46 |
SELECT * FROM articles WHERE |
| 47 |
MATCH(title, body) AGAINST('数据库' IN BOOLEAN MODE); |
| 48 |
|
| 49 |
--sorted_result |
| 50 |
SELECT * FROM articles WHERE |
| 51 |
MATCH(title, body) AGAINST('"数据库"' IN BOOLEAN MODE); |
| 52 |
|
| 53 |
DROP TABLE articles; |
| 54 |
|
| 55 |
-- echo # Test Case 2: Test Chinese with GBK Charset. |
| 56 |
CREATE TABLE articles ( |
| 57 |
id INT UNSIGNED AUTO_INCREMENT NOT NULL PRIMARY KEY, |
| 58 |
title VARCHAR(200), |
| 59 |
body TEXT |
| 60 |
) ENGINE=InnoDB DEFAULT CHARACTER SET gbk COLLATE gbk_chinese_ci; |
| 61 |
|
| 62 |
INSERT INTO articles (title, body) VALUES |
| 63 |
('数据库是数据的结构化集合','它可以是任何东西'), |
| 64 |
('从简单的购物清单到画展','或企业网络中的海量数据'); |
| 65 |
|
| 66 |
ALTER TABLE articles ADD FULLTEXT INDEX (title, body) WITH PARSER ngram; |
| 67 |
|
| 68 |
INSERT INTO articles (title, body) VALUES |
| 69 |
('要想将数据添加到数据库','或访问、处理计算机数据库中保存的数据'), |
| 70 |
('需要使用数据库管理系统','计算机是处理大量数据的理想工具'); |
| 71 |
|
| 72 |
--sorted_result |
| 73 |
SELECT * FROM articles WHERE |
| 74 |
MATCH(title, body) AGAINST('数据'); |
| 75 |
|
| 76 |
--sorted_result |
| 77 |
SELECT * FROM articles WHERE |
| 78 |
MATCH(title, body) AGAINST('数*' IN BOOLEAN MODE); |
| 79 |
|
| 80 |
SELECT * FROM articles WHERE |
| 81 |
MATCH(title, body) AGAINST('+数据库 -计算机' IN BOOLEAN MODE); |
| 82 |
|
| 83 |
SELECT * FROM articles WHERE |
| 84 |
MATCH(title, body) AGAINST('+(数据库 处理) -计算机' IN BOOLEAN MODE); |
| 85 |
|
| 86 |
--sorted_result |
| 87 |
SELECT * FROM articles WHERE |
| 88 |
MATCH(title, body) AGAINST('数据库'); |
| 89 |
|
| 90 |
--sorted_result |
| 91 |
SELECT * FROM articles WHERE |
| 92 |
MATCH(title, body) AGAINST('数据库' IN BOOLEAN MODE); |
| 93 |
|
| 94 |
--sorted_result |
| 95 |
SELECT * FROM articles WHERE |
| 96 |
MATCH(title, body) AGAINST('"数据库"' IN BOOLEAN MODE); |
| 97 |
|
| 98 |
DROP TABLE articles; |
| 99 |
|
| 100 |
-- echo # Test Case 3: Test Chinese with UTF8 Charset. |
| 101 |
CREATE TABLE articles ( |
| 102 |
id INT UNSIGNED AUTO_INCREMENT NOT NULL PRIMARY KEY, |
| 103 |
title VARCHAR(200), |
| 104 |
body TEXT |
| 105 |
) ENGINE=InnoDB DEFAULT CHARACTER SET utf8mb4 COLLATE utf8mb4_general_ci; |
| 106 |
|
| 107 |
INSERT INTO articles (title, body) VALUES |
| 108 |
('数据库是数据的结构化集合','它可以是任何东西'), |
| 109 |
('从简单的购物清单到画展','或企业网络中的海量数据'); |
| 110 |
|
| 111 |
ALTER TABLE articles ADD FULLTEXT INDEX (title, body) WITH PARSER ngram; |
| 112 |
|
| 113 |
INSERT INTO articles (title, body) VALUES |
| 114 |
('要想将数据添加到数据库','或访问、处理计算机数据库中保存的数据'), |
| 115 |
('需要使用数据库管理系统','计算机是处理大量数据的理想工具'); |
| 116 |
|
| 117 |
--sorted_result |
| 118 |
SELECT * FROM articles WHERE |
| 119 |
MATCH(title, body) AGAINST('数据'); |
| 120 |
|
| 121 |
--sorted_result |
| 122 |
SELECT * FROM articles WHERE |
| 123 |
MATCH(title, body) AGAINST('数*' IN BOOLEAN MODE); |
| 124 |
|
| 125 |
SELECT * FROM articles WHERE |
| 126 |
MATCH(title, body) AGAINST('+数据库 -计算机' IN BOOLEAN MODE); |
| 127 |
|
| 128 |
SELECT * FROM articles WHERE |
| 129 |
MATCH(title, body) AGAINST('+(数据库 处理) -计算机' IN BOOLEAN MODE); |
| 130 |
|
| 131 |
--sorted_result |
| 132 |
SELECT * FROM articles WHERE |
| 133 |
MATCH(title, body) AGAINST('数据库'); |
| 134 |
|
| 135 |
--sorted_result |
| 136 |
SELECT * FROM articles WHERE |
| 137 |
MATCH(title, body) AGAINST('数据库' IN BOOLEAN MODE); |
| 138 |
|
| 139 |
--sorted_result |
| 140 |
SELECT * FROM articles WHERE |
| 141 |
MATCH(title, body) AGAINST('"数据库"' IN BOOLEAN MODE); |
| 142 |
|
| 143 |
DROP TABLE articles; |
| 144 |
|
| 145 |
-- echo # Test Case 4: Test Chinese with gb18030 Charset. |
| 146 |
CREATE TABLE articles ( |
| 147 |
id INT UNSIGNED AUTO_INCREMENT NOT NULL PRIMARY KEY, |
| 148 |
title VARCHAR(200), |
| 149 |
body TEXT |
| 150 |
) ENGINE=InnoDB DEFAULT CHARSET=gb18030 COLLATE=gb18030_chinese_ci; |
| 151 |
|
| 152 |
|
| 153 |
INSERT INTO articles (title, body) VALUES |
| 154 |
('数据库是数据的结构化集合','它可以是任何东西'), |
| 155 |
('从简单的购物清单到画展','或企业网络中的海量数据'); |
| 156 |
|
| 157 |
ALTER TABLE articles ADD FULLTEXT INDEX `idx1` (title, body) WITH PARSER ngram; |
| 158 |
|
| 159 |
INSERT INTO articles (title, body) VALUES |
| 160 |
('要想将数据添加到数据库','或访问、处理计算机数据库中保存的数据'), |
| 161 |
('需要使用数据库管理系统','计算机是处理大量数据的理想工具'); |
| 162 |
|
| 163 |
--sorted_result |
| 164 |
SELECT * FROM articles WHERE |
| 165 |
MATCH(title, body) AGAINST('数据'); |
| 166 |
|
| 167 |
--sorted_result |
| 168 |
SELECT * FROM articles WHERE |
| 169 |
MATCH(title, body) AGAINST('数*' IN BOOLEAN MODE); |
| 170 |
|
| 171 |
SELECT * FROM articles WHERE |
| 172 |
MATCH(title, body) AGAINST('+数据库 -计算机' IN BOOLEAN MODE); |
| 173 |
|
| 174 |
SELECT * FROM articles WHERE |
| 175 |
MATCH(title, body) AGAINST('+(数据库 处理) -计算机' IN BOOLEAN MODE); |
| 176 |
|
| 177 |
--sorted_result |
| 178 |
SELECT * FROM articles WHERE |
| 179 |
MATCH(title, body) AGAINST('数据库'); |
| 180 |
|
| 181 |
--sorted_result |
| 182 |
SELECT * FROM articles WHERE |
| 183 |
MATCH(title, body) AGAINST('数据库' IN BOOLEAN MODE); |
| 184 |
|
| 185 |
--sorted_result |
| 186 |
SELECT * FROM articles WHERE |
| 187 |
MATCH(title, body) AGAINST('"数据库"' IN BOOLEAN MODE); |
| 188 |
|
| 189 |
--sorted_result |
| 190 |
SELECT * FROM articles WHERE |
| 191 |
MATCH(title, body) AGAINST('数据库' WITH QUERY EXPANSION); |
| 192 |
|
| 193 |
ALTER TABLE articles DROP INDEX `idx1`; |
| 194 |
SELECT * FROM articles; |
| 195 |
DROP TABLE articles; |
| 196 |
|
| 197 |
|
| 198 |
-- echo # Test Case 5: Test Japanese with UJIS Charset. |
| 199 |
CREATE TABLE articles ( |
| 200 |
id INT UNSIGNED AUTO_INCREMENT NOT NULL PRIMARY KEY, |
| 201 |
title VARCHAR(200), |
| 202 |
body TEXT |
| 203 |
) ENGINE=InnoDB DEFAULT CHARACTER SET ujis COLLATE ujis_japanese_ci; |
| 204 |
|
| 205 |
INSERT INTO articles (title, body) VALUES |
| 206 |
('データベースは、構造化されたコレクションです','それは何もすることができます'), |
| 207 |
('シンプルな買い物リストから画像ギャラリーへ','または企業ネットワーク内のデータの膨大な量'); |
| 208 |
|
| 209 |
ALTER TABLE articles ADD FULLTEXT INDEX (title, body) WITH PARSER ngram; |
| 210 |
|
| 211 |
INSERT INTO articles (title, body) VALUES |
| 212 |
('データベースにデータを追加するには','コンピュータのデータベースに格納されているアクセス、およびプロセスデータ'), |
| 213 |
('データベース管理システムを使用する必要が','コンピュータは、大量のデータを処理するための理想的なツールである'); |
| 214 |
|
| 215 |
--sorted_result |
| 216 |
SELECT * FROM articles WHERE |
| 217 |
MATCH(title, body) AGAINST('データ'); |
| 218 |
|
| 219 |
--sorted_result |
| 220 |
SELECT * FROM articles WHERE |
| 221 |
MATCH(title, body) AGAINST('デ*' IN BOOLEAN MODE); |
| 222 |
|
| 223 |
SELECT * FROM articles WHERE |
| 224 |
MATCH(title, body) AGAINST('+データベース -コンピュータ' IN BOOLEAN MODE); |
| 225 |
|
| 226 |
SELECT * FROM articles WHERE |
| 227 |
MATCH(title, body) AGAINST('+(データベース 処理) -コンピュータ' IN BOOLEAN MODE); |
| 228 |
|
| 229 |
--sorted_result |
| 230 |
SELECT * FROM articles WHERE |
| 231 |
MATCH(title, body) AGAINST('データベース'); |
| 232 |
|
| 233 |
--sorted_result |
| 234 |
SELECT * FROM articles WHERE |
| 235 |
MATCH(title, body) AGAINST('データベース' IN BOOLEAN MODE); |
| 236 |
|
| 237 |
--sorted_result |
| 238 |
SELECT * FROM articles WHERE |
| 239 |
MATCH(title, body) AGAINST('"データベース"' IN BOOLEAN MODE); |
| 240 |
|
| 241 |
DROP TABLE articles; |
| 242 |
|
| 243 |
-- echo # Test Case 6: Test Korean with EUCKR Charset. |
| 244 |
CREATE TABLE articles ( |
| 245 |
id INT UNSIGNED AUTO_INCREMENT NOT NULL PRIMARY KEY, |
| 246 |
title VARCHAR(200), |
| 247 |
body TEXT |
| 248 |
) ENGINE=InnoDB DEFAULT CHARACTER SET euckr COLLATE euckr_korean_ci; |
| 249 |
|
| 250 |
INSERT INTO articles (title, body) VALUES |
| 251 |
('데이터베이스의 구조 모음입니다','그것은 무엇이든 될 수 있습니다'), |
| 252 |
('간단한 쇼핑 목록에서 사진 갤러리','또는 기업 네트워크에서 데이터의 엄청난 양의'); |
| 253 |
|
| 254 |
ALTER TABLE articles ADD FULLTEXT INDEX (title, body) WITH PARSER ngram; |
| 255 |
|
| 256 |
INSERT INTO articles (title, body) VALUES |
| 257 |
('데이터베이스에 데이터를 추가하려면','액세스, 프로세스 데이터를 컴퓨터 데이터베이스에 저장'), |
| 258 |
('데이터베이스 관리 시스템을 사용해야합니다','컴퓨터는 많은 양의 데이터를 처리하기위한 이상적인 도구입니다'); |
| 259 |
|
| 260 |
--sorted_result |
| 261 |
SELECT * FROM articles WHERE |
| 262 |
MATCH(title, body) AGAINST('데이터'); |
| 263 |
|
| 264 |
--sorted_result |
| 265 |
SELECT * FROM articles WHERE |
| 266 |
MATCH(title, body) AGAINST('데*' IN BOOLEAN MODE); |
| 267 |
|
| 268 |
SELECT * FROM articles WHERE |
| 269 |
MATCH(title, body) AGAINST('+데이터베이스 -컴퓨터' IN BOOLEAN MODE); |
| 270 |
|
| 271 |
SELECT * FROM articles WHERE |
| 272 |
MATCH(title, body) AGAINST('+(데이터베이스 처리) -컴퓨터' IN BOOLEAN MODE); |
| 273 |
|
| 274 |
--sorted_result |
| 275 |
SELECT * FROM articles WHERE |
| 276 |
MATCH(title, body) AGAINST('데이터베이스'); |
| 277 |
|
| 278 |
--sorted_result |
| 279 |
SELECT * FROM articles WHERE |
| 280 |
MATCH(title, body) AGAINST('데이터베이스' IN BOOLEAN MODE); |
| 281 |
|
| 282 |
--sorted_result |
| 283 |
SELECT * FROM articles WHERE |
| 284 |
MATCH(title, body) AGAINST('"데이터베이스"' IN BOOLEAN MODE); |
| 285 |
|
| 286 |
DROP TABLE articles; |
| 287 |
|
| 288 |
-- echo # Test Case 7: Test different ngram token size. |
| 289 |
CREATE TABLE articles ( |
| 290 |
id INT UNSIGNED AUTO_INCREMENT NOT NULL PRIMARY KEY, |
| 291 |
title VARCHAR(200), |
| 292 |
body TEXT |
| 293 |
) ENGINE=InnoDB DEFAULT CHARACTER SET gb2312 COLLATE gb2312_chinese_ci; |
| 294 |
|
| 295 |
INSERT INTO articles (title, body) VALUES |
| 296 |
('数据库是数据的结构化集合','它可以是任何东西'), |
| 297 |
('从简单的购物清单到画展','或企业网络中的海量数据'), |
| 298 |
('要想将数据添加到数据库','或访问、处理计算机数据库中保存的数据'), |
| 299 |
('需要使用数据库管理系统','计算机是处理大量数据的理想工具'); |
| 300 |
|
| 301 |
# bigram |
| 302 |
CREATE FULLTEXT INDEX ft_index ON articles(title, body) WITH PARSER ngram; |
| 303 |
|
| 304 |
# unigram |
| 305 |
let $ngram_token_size=1; |
| 306 |
-- source suite/innodb_fts/include/ngram_token_size.inc |
| 307 |
|
| 308 |
# trigram |
| 309 |
let $ngram_token_size=3; |
| 310 |
-- source suite/innodb_fts/include/ngram_token_size.inc |
| 311 |
|
| 312 |
# ngram = 5 |
| 313 |
let $ngram_token_size=5; |
| 314 |
-- source suite/innodb_fts/include/ngram_token_size.inc |
| 315 |
|
| 316 |
# ngram = 7 |
| 317 |
let $ngram_token_size=7; |
| 318 |
-- source suite/innodb_fts/include/ngram_token_size.inc |
| 319 |
|
| 320 |
# Restore to bigram |
| 321 |
let $ngram_token_size=2; |
| 322 |
-- source suite/innodb_fts/include/ngram_token_size.inc |
| 323 |
|
| 324 |
DROP TABLE articles; |
| 325 |
|
| 326 |
-- echo # Test Case 8: Test Optimize Table. |
| 327 |
CREATE TABLE articles ( |
| 328 |
id INT UNSIGNED AUTO_INCREMENT NOT NULL PRIMARY KEY, |
| 329 |
title VARCHAR(200), |
| 330 |
body TEXT |
| 331 |
) ENGINE=InnoDB DEFAULT CHARACTER SET gb2312 COLLATE gb2312_chinese_ci; |
| 332 |
|
| 333 |
INSERT INTO articles (title, body) VALUES |
| 334 |
('数据库是数据的结构化集合','它可以是任何东西'), |
| 335 |
('从简单的购物清单到画展','或企业网络中的海量数据'); |
| 336 |
|
| 337 |
CREATE FULLTEXT INDEX ft_index ON articles(title, body) WITH PARSER ngram; |
| 338 |
|
| 339 |
SET GLOBAL innodb_ft_aux_table="test/articles"; |
| 340 |
|
| 341 |
SELECT * FROM INFORMATION_SCHEMA.INNODB_FT_INDEX_TABLE; |
| 342 |
|
| 343 |
SELECT * FROM articles WHERE |
| 344 |
MATCH(title, body) AGAINST('数据'); |
| 345 |
|
| 346 |
DELETE FROM articles WHERE id = 1; |
| 347 |
|
| 348 |
SET GLOBAL INNODB_OPTIMIZE_FULLTEXT_ONLY=1; |
| 349 |
OPTIMIZE TABLE articles; |
| 350 |
SET GLOBAL INNODB_OPTIMIZE_FULLTEXT_ONLY=default; |
| 351 |
|
| 352 |
SELECT * FROM INFORMATION_SCHEMA.INNODB_FT_INDEX_TABLE; |
| 353 |
|
| 354 |
SELECT * FROM articles WHERE |
| 355 |
MATCH(title, body) AGAINST('数据'); |
| 356 |
|
| 357 |
SET GLOBAL innodb_ft_aux_table=default; |
| 358 |
|
| 359 |
DROP TABLE articles; |
| 360 |
|
| 361 |
-- echo # Test Case 9: Test Rename Index |
| 362 |
--replace_regex /\.dll/.so/ |
| 363 |
eval INSTALL PLUGIN simple_parser SONAME '$SIMPLE_PARSER'; |
| 364 |
|
| 365 |
CREATE TABLE articles ( |
| 366 |
id INT UNSIGNED AUTO_INCREMENT NOT NULL PRIMARY KEY, |
| 367 |
title VARCHAR(200), |
| 368 |
body TEXT |
| 369 |
) ENGINE=InnoDB DEFAULT CHARACTER SET gb2312 COLLATE gb2312_chinese_ci; |
| 370 |
|
| 371 |
INSERT INTO articles (title, body) VALUES ('数据的结合','数据库'); |
| 372 |
|
| 373 |
CREATE FULLTEXT INDEX ft_1 ON articles(title) WITH PARSER ngram; |
| 374 |
|
| 375 |
SELECT * FROM articles WHERE MATCH(title) AGAINST('数据'); |
| 376 |
|
| 377 |
ALTER TABLE articles RENAME INDEX ft_1 TO ft_2; |
| 378 |
|
| 379 |
INSERT INTO articles (title, body) VALUES ('数据的结合','数据库'); |
| 380 |
|
| 381 |
SELECT * FROM articles WHERE MATCH(title) AGAINST('数据'); |
| 382 |
|
| 383 |
ALTER TABLE articles RENAME INDEX ft_2 to ft_1, |
| 384 |
ADD FULLTEXT INDEX ft_2(body) WITH PARSER simple_parser; |
| 385 |
|
| 386 |
INSERT INTO articles (title, body) VALUES ('数据的结合','数据库'); |
| 387 |
|
| 388 |
SELECT * FROM articles WHERE MATCH(title) AGAINST('数据'); |
| 389 |
|
| 390 |
SELECT * FROM articles WHERE MATCH(body) AGAINST('数据'); |
| 391 |
SELECT * FROM articles WHERE MATCH(body) AGAINST('数据库'); |
| 392 |
|
| 393 |
ALTER TABLE articles RENAME INDEX ft_2 to ft_1, RENAME INDEX ft_1 to ft_2; |
| 394 |
|
| 395 |
INSERT INTO articles (title, body) VALUES ('数据的结合','数据库'); |
| 396 |
|
| 397 |
SELECT * FROM articles WHERE MATCH(title) AGAINST('数据'); |
| 398 |
|
| 399 |
SELECT * FROM articles WHERE MATCH(body) AGAINST('数据'); |
| 400 |
SELECT * FROM articles WHERE MATCH(body) AGAINST('数据库'); |
| 401 |
|
| 402 |
DROP TABLE articles; |
| 403 |
|
| 404 |
UNINSTALL PLUGIN simple_parser; |
| 405 |
|
| 406 |
-- echo # Test Case 10: Test Fulltext with UCS2 |
| 407 |
CREATE TABLE articles ( |
| 408 |
id INT UNSIGNED AUTO_INCREMENT NOT NULL PRIMARY KEY, |
| 409 |
title VARCHAR(200), |
| 410 |
body TEXT |
| 411 |
) ENGINE=InnoDB DEFAULT CHARACTER SET ucs2; |
| 412 |
|
| 413 |
INSERT INTO articles (title, body) VALUES ('数据的结合','数据库'); |
| 414 |
|
| 415 |
# Fulltext with charset UCS2 is not allowed. |
| 416 |
-- error ER_BAD_FT_COLUMN |
| 417 |
CREATE FULLTEXT INDEX ft_idx ON articles(title) WITH PARSER ngram; |
| 418 |
|
| 419 |
-- error ER_BAD_FT_COLUMN |
| 420 |
CREATE FULLTEXT INDEX ft_idx ON articles(title); |
| 421 |
|
| 422 |
DROP TABLE articles; |