mirror of
https://github.com/ClickHouse/ClickHouse.git
synced 2024-11-19 06:01:57 +00:00
2c828338f4
This commit migrates ClickHouse to Vectorscan. The first 10 min of [0] explain the reasons for it. (*) Addresses (but does not resolve) #38046 (*) Config parameter names (e.g. "max_hyperscan_regexp_length") are preserved for compatibility. Likewise, error codes (e.g. "ErrorCodes::HYPERSCAN_CANNOT_SCAN_TEXT") and function/class names (e.g. "HyperscanDeleter") are preserved as vectorscan aims to be a drop-in replacement. [0] https://www.youtube.com/watch?v=KlZWmmflW6M
95 lines
13 KiB
SQL
95 lines
13 KiB
SQL
-- Tags: no-fasttest, use-vectorscan
|
||
|
||
select 0 = multiMatchAny(materialize('mpnsguhwsitzvuleiwebwjfitmsg'), ['wbirxqoabpblrnvvmjizj', 'cfcxhuvrexyzyjsh', 'oldhtubemyuqlqbwvwwkwin', 'bumoozxdkjglzu', 'intxlfohlxmajjomw', 'dxkeghohv', 'arsvmwwkjeopnlwnan', 'ouugllgowpqtaxslcopkytbfhifaxbgt', 'hkedmjlbcrzvryaopjqdjjc', 'tbqkljywstuahzh', 'o', 'wowoclosyfcuwotmvjygzuzhrery', 'vpefjiffkhlggntcu', 'ytdixvasrorhripzfhjdmlhqksmctyycwp']) from system.numbers limit 10;
|
||
select 0 = multiMatchAny(materialize('qjjzqexjpgkglgxpzrbqbnskq'), ['vaiatcjacmlffdzsejpdareqzy', 'xspcfzdufkmecud', 'bcvtbuqtctq', 'nkcopwbfytgemkqcfnnno', 'dylxnzuyhq', 'tno', 'scukuhufly', 'cdyquzuqlptv', 'ohluyfeksyxepezdhqmtfmgkvzsyph', 'ualzwtahvqvtijwp', 'jg', 'gwbawqlngzcknzgtmlj', 'qimvjcgbkkp', 'eaedbcgyrdvv', 'qcwrncjoewwedyyewcdkh', 'uqcvhngoqngmitjfxpznqomertqnqcveoqk', 'ydrgjiankgygpm', 'axepgap']) from system.numbers limit 10;
|
||
select 0 = multiMatchAny(materialize('fdkmtqmxnegwvnjhghjq'), ['vynkybvdmhgeezybbdqfrukibisj', 'knazzamgjjpavwhvdkwigykh', 'peumnifrmdhhmrqqnemw', 'lmsnyvqoisinlaqobxojlwfbi', 'oqwfzs', 'dymudxxeodwjpgbibnkvr', 'vomtfsnizkplgzktqyoiw', 'yoyfuhlpgrzds', 'cefao', 'gi', 'srpgxfjwl', 'etsjusdeiwbfe', 'ikvtzdopxo', 'ljfkavrau', 'soqdhxtenfrkmeic', 'ktprjwfcelzbup', 'pcvuoddqwsaurcqdtjfnczekwni', 'agkqkqxkfbkfgyqliahsljim']) from system.numbers limit 10;
|
||
select 1 = multiMatchAny(materialize('khljxzxlpcrxpkrfybbfk'), ['', 'lpc', 'rxpkrfybb', 'crxp', '', 'pkr', 'jxzxlpcrxpkrf', '', 'xzxlpcr', 'xpk', 'fyb', 'xzxlpcrxpkrfybbfk', 'k', 'lpcrxp', 'ljxzxlpcr', 'r', 'pkr', 'fk']) from system.numbers limit 10;
|
||
select 0 = multiMatchAny(materialize('rbrizgjbigvzfnpgmpkqxoqxvdj'), ['ee', 'cohqnb', 'msol', 'yhlujcvhklnhuomy', 'ietn', 'vgmnlkcsybtokrepzrm', 'wspiryefojxysgrzsxyrluykxfnnbzdstcel', 'mxisnsivndbefqxwznimwgazuulupbaihavg', 'vpzdjvqqeizascxmzdhuq', 'pgvncohlxcqjhfkm', 'mbaypcnfapltsegquurahlsruqvipfhrhq', 'ioxjbcyyqujfveujfhnfdfokfcrlsincjbdt', 'cnvlujyowompdrqjwjx', 'wobwed', 'kdfhaoxiuifotmptcmdbk', 'leoamsnorcvtlmokdomkzuo', 'jjw', 'ogugysetxuqmvggneosbsfbonszepsatq']) from system.numbers limit 10;
|
||
select 0 = multiMatchAny(materialize('uymwxzyjbfegbhgswiqhinf'), ['lizxzbzlwljkr', 'ukxygktlpzuyijcqeqktxenlaqi', 'onperabgbdiafsxwbvpjtyt', 'xfqgoqvhqph', 'aflmcwabtwgmajmmqelxwkaolyyhmdlc', 'yfz', 'meffuiaicvwed', 'hhzvgmifzamgftkifaeowayjrnnzw', 'nwewybtajv', 'ectiye', 'epjeiljegmqqjncubj', 'zsjgftqjrn', 'pssng', 'raqoarfhdoeujulvqmdo']) from system.numbers limit 10;
|
||
select 0 = multiMatchAny(materialize('omgghgnzjmecpzqmtcvw'), ['fjhlzbszodmzavzg', 'gfofrnwrxprkfiokv', 'jmjiiqpgznlmyrxwewzqzbe', 'pkyrsqkltlmxr', 'crqgkgqkkyujcyoc', 'endagbcxwqhueczuasykmajfsvtcmh', 'xytmxtrnkdysuwltqomehddp', 'etmdxyyfotfyifwvbykghijvwv', 'mwqtgrncyhkfhjdg', 'iuvymofrqpp', 'pgllsdanlhzqhkstwsmzzftp', 'disjylcceufxtjdvhy']) from system.numbers limit 10;
|
||
select 1 = multiMatchAny(materialize('mznihnmshftvnmmhnrulizzpslq'), ['nrul', 'mshftvnmmhnr', 'z', 'mhnrulizzps', 'hftvnmmhnrul', 'ihnmshftvnmmhnrulizzp', 'izz', '', 'uli', 'nihnmshftvnmmhnru', 'hnrulizzp', 'nrulizz']) from system.numbers limit 10;
|
||
select 1 = multiMatchAny(materialize('ruqmqrsxrbftvruvahonradau'), ['uqmqrsxrbft', 'ftv', 'tvruvahonrad', 'mqrsxrbftvruvahon', 'rbftvruvah', 'qrsxrbftvru', 'o', 'ahonradau', 'a', 'ft', '', 'u', 'rsxrbftvruvahonradau', 'ruvahon', 'bftvruvahonradau', 'qrsxrbftvru', 't', 'vahonrada', 'vruvahonradau', 'onra']) from system.numbers limit 10;
|
||
select 1 = multiMatchAny(materialize('gpsevxtcoeexrltyzduyidmtzxf'), ['exrltyzduyid', 'vxtcoeexrltyz', 'xr', 'ltyzduyidmt', 'yzduy', 'exr', 'coeexrltyzduy', 'coeexrltyzduy', 'rlty', 'rltyzduyidm', 'exrltyz', 'xtcoeexrlty', 'vxtcoeexrltyzduyidm', '', 'coeexrl', 'sevxtcoeexrltyzdu', 'dmt', '']) from system.numbers limit 10;
|
||
select 0 = multiMatchAny(materialize('dyhycfhzyewaikgursyxfkuv'), ['sktnofpugrmyxmbizzrivmhn', 'fhlgadpoqcvktbfzncxbllvwutdawmw', 'eewzjpcgzrqmltbgmhafwlwqb', 'tpogbkyj', 'rtllntxjgkzs', 'mirbvsqexscnzglogigbujgdwjvcv', 'iktwpgjsakemewmahgqza', 'xgfvzkvqgiuoihjjnxwwpznxhz', 'nxaumpaknreklbwynvxdsmatjekdlxvklh', 'zadzwqhgfxqllihuudozxeixyokhny', 'tdqpgfpzexlkslodps', 'slztannufxaabqfcjyfquafgfhfb', 'xvjldhfuwurvkb', 'aecv', 'uycfsughpikqsbcmwvqygdyexkcykhbnau', 'jr']) from system.numbers limit 10;
|
||
select 1 = multiMatchAny(materialize('vbcsettndwuntnruiyclvvwoo'), ['dwuntnru', '', 'ttndwuntnruiyclvv', 'ntnr', 'nruiyclvvw', 'wo', '', 'bcsettndwuntnruiycl', 'yc', 'untnruiyclvvw', 'csettndwuntnr', 'ntnruiyclvvwo']) from system.numbers limit 10;
|
||
select 0 = multiMatchAny(materialize('pqqnugshlczcuxhpjxjbcnro'), ['dpeedqy', 'rtsc', 'jdgla', 'qkgudqjiyzvlvsj', 'xmfxawhijgxxtydbd', 'ebgzazqthb', 'wyrjhvhwzhmpybnylirrn', 'iviqbyuclayqketooztwegtkgwnsezfl', 'bhvidy', 'hijctxxweboq', 't', 'osnzfbziidteiaifgaanm']) from system.numbers limit 10;
|
||
select 1 = multiMatchAny(materialize('loqchlxspwuvvccucskuytr'), ['', 'k', 'qchlxspwu', 'u', 'hlxspwuvv', 'wuvvccucsku', 'vcc', 'uyt', 'uvv', 'spwu', 'ytr', 'wuvvccucs', 'xspwuv', 'lxspwuvvccuc', 'spwuvvccu', 'oqchlxspwuvvccucskuy']) from system.numbers limit 10;
|
||
select 1 = multiMatchAny(materialize('pjjyzupzwllshlnatiujmwvaofr'), ['lnatiujmwvao', '', 'zupzwllsh', 'nati', 'wllshl', 'hlnatiujmwv', 'mwvao', 'shlnat', 'ati', 'wllshlnatiujmwvao', 'wllshlnatiujmwvaofr', 'nat']) from system.numbers limit 10;
|
||
select 1 = multiMatchAny(materialize('iketunkleyaqaxdlocci'), ['nkleyaqaxd', 'etunkleyaq', 'yaqaxdlocci', 'tunkleyaq', 'eyaqaxdlocc', 'leyaq', 'nkleyaqaxdl', 'tunkleya', 'kleyaqa', 'etunkleya', 'leyaqa', 'dlo', 'yaqa', 'leyaqaxd', 'etunkleyaq', '']) from system.numbers limit 10;
|
||
select 0 = multiMatchAny(materialize('drqianqtangmgbdwruvblkqd'), ['wusajejyucamkyl', 'wsgibljugzrpkniliy', 'lhwqqiuafwffyersqjgjvvvfurx', 'jfokpzzxfdonelorqu', 'ccwkpcgac', 'jmyulqpndkmzbfztobwtm', 'rwrgfkccgxht', 'ggldjecrgbngkonphtcxrkcviujihidjx', 'spwweavbiokizv', 'lv', 'krb', 'vstnhvkbwlqbconaxgbfobqky', 'pvxwdc', 'thrl', 'ahsblffdveamceonqwrbeyxzccmux', 'yozji', 'oejtaxwmeovtqtz', 'zsnzznvqpxdvdxhznxrjn', 'hse', 'kcmkrccxmljzizracxwmpoaggywhdfpxkq']) from system.numbers limit 10;
|
||
select 0 = multiMatchAny(materialize('yasnpckniistxcejowfijjsvkdajz'), ['slkpxhtsmrtvtm', 'crsbq', 'rdeshtxbfrlfwpsqojassxmvlfbzefldavmgme', 'ipetilcbpsfroefkjirquciwtxhrimbmwnlyv', 'knjpwkmdwbvdbapuyqbtsw', 'horueidziztxovqhsicnklmharuxhtgrsr', 'ofohrgpz', 'oneqnwyevbaqsonrcpmxcynflojmsnix', 'shg', 'nglqzczevgevwawdfperpeytuodjlf']) from system.numbers limit 10;
|
||
select 0 = multiMatchAny(materialize('ueptpscfgxhplwsueckkxs'), ['ohhygchclbpcdwmftperprn', 'dvpjdqmqckekndvcerqrpkxen', 'lohhvarnmyi', 'zppd', 'qmqxgfewitsunbuhffozcpjtc', 'hsjbioisycsrawktqssjovkmltxodjgv', 'dbzuunwbkrtosyvctdujqtvaawfnvuq', 'gupbvpqthqxae', 'abjdmijaaiasnccgxttmqdsz', 'uccyumqoyqe', 'kxxliepyzlc', 'wbqcqtbyyjbqcgdbpkmzugksmcxhvr', 'piedxm', 'uncpphzoif', 'exkdankwck', 'qeitzozdrqopsergzr', 'hesgrhaftgesnzflrrtjdobxhbepjoas', 'wfpexx']) from system.numbers limit 10;
|
||
select 1 = multiMatchAny(materialize('ldrzgttlqaphekkkdukgngl'), ['gttlqaphekkkdukgn', 'ekkkd', 'gttlqaphe', 'qaphek', 'h', 'kdu', 'he', 'phek', '', 'drzgttlqaphekkkd']) from system.numbers limit 10;
|
||
select 1 = multiMatchAny(materialize('ololo'), ['ololo', 'ololo', 'ololo']);
|
||
SELECT 1 = multiMatchAny(materialize('khljxzxlpcrxpkrfybbfk'), ['k']);
|
||
|
||
select 1 = multiMatchAny(materialize(''), ['']);
|
||
select 0 = multiMatchAny(materialize(''), ['some string']);
|
||
select 1 = multiMatchAny(materialize('abc'), ['']);
|
||
select 1 = multiMatchAny(materialize('abc'), ['']) from system.numbers limit 10;
|
||
|
||
select 0 = multiMatchAny(materialize('abc'), ['defgh']);
|
||
select 0 = multiMatchAny(materialize('abc'), ['defg']);
|
||
select 0 = multiMatchAny(materialize('abc'), ['def']);
|
||
select 0 = multiMatchAny(materialize('abc'), ['de']);
|
||
select 0 = multiMatchAny(materialize('abc'), ['d']);
|
||
|
||
|
||
select 1 = multiMatchAny(materialize('abc'), ['b']) from system.numbers limit 10;
|
||
select 1 = multiMatchAny(materialize('abc'), ['bc']) from system.numbers limit 10;
|
||
select 0 = multiMatchAny(materialize('abc'), ['bcde']) from system.numbers limit 10;
|
||
select 0 = multiMatchAny(materialize('abc'), ['bcdef']) from system.numbers limit 10;
|
||
select 0 = multiMatchAny(materialize('abc'), ['bcdefg']) from system.numbers limit 10;
|
||
select 0 = multiMatchAny(materialize('abc'), ['bcdefgh']) from system.numbers limit 10;
|
||
|
||
|
||
select 0 = multiMatchAny(materialize('abc'), ['abcdefg']) from system.numbers limit 10;
|
||
select 0 = multiMatchAny(materialize('abc'), ['abcdef']) from system.numbers limit 10;
|
||
select 0 = multiMatchAny(materialize('abc'), ['abcde']) from system.numbers limit 10;
|
||
select 0 = multiMatchAny(materialize('abc'), ['abcd']) from system.numbers limit 10;
|
||
select 1 = multiMatchAny(materialize('abc'), ['abc']) from system.numbers limit 10;
|
||
select 1 = multiMatchAny(materialize('abc'), ['ab']) from system.numbers limit 10;
|
||
select 1 = multiMatchAny(materialize('abc'), ['a']) from system.numbers limit 10;
|
||
|
||
select 1 = multiMatchAny(materialize('abcd'), ['c']) from system.numbers limit 10;
|
||
select 1 = multiMatchAny(materialize('abcd'), ['cd']) from system.numbers limit 10;
|
||
select 0 = multiMatchAny(materialize('abcd'), ['cde']) from system.numbers limit 10;
|
||
select 0 = multiMatchAny(materialize('abcd'), ['cdef']) from system.numbers limit 10;
|
||
select 0 = multiMatchAny(materialize('abcd'), ['cdefg']) from system.numbers limit 10;
|
||
select 0 = multiMatchAny(materialize('abcd'), ['cdefgh']) from system.numbers limit 10;
|
||
|
||
select 0 = multiMatchAny(materialize('abc'), ['defgh']) from system.numbers limit 10;
|
||
select 0 = multiMatchAny(materialize('abc'), ['defg']) from system.numbers limit 10;
|
||
select 0 = multiMatchAny(materialize('abc'), ['def']) from system.numbers limit 10;
|
||
select 0 = multiMatchAny(materialize('abc'), ['de']) from system.numbers limit 10;
|
||
select 0 = multiMatchAny(materialize('abc'), ['d']) from system.numbers limit 10;
|
||
|
||
select 1 = multiMatchAny(materialize('abc'), ['...']) from system.numbers limit 10;
|
||
select 1 = multiMatchAny(materialize('a\nbc'), ['a?bc']) from system.numbers limit 10;
|
||
select 1 = multiMatchAny(materialize('a\nbc'), ['a.bc']) from system.numbers limit 10;
|
||
select 1 = multiMatchAny(materialize('a\0bc'), ['a?bc']) from system.numbers limit 10;
|
||
select 1 = multiMatchAny(materialize('a\0bc'), ['a.bc']) from system.numbers limit 10;
|
||
select 1 = multiMatchAny(materialize('abcdef'), ['a.....']) from system.numbers limit 10;
|
||
select 0 = multiMatchAny(materialize('abcdef'), ['a......']) from system.numbers limit 10;
|
||
select 1 = multiMatchAny(materialize('abcdef'), ['a......', 'a.....']) from system.numbers limit 10;
|
||
select 0 = multiMatchAny(materialize('aaaa'), ['.*aa.*aaa.*', 'aaaaaa{2}', '\(aa\){3}']) from system.numbers limit 10;
|
||
select 1 = multiMatchAny(materialize('abc'), ['a\0d']) from system.numbers limit 10;
|
||
|
||
select 1 = multiMatchAnyIndex(materialize('gogleuedeuniangoogle'), ['google', 'unian1']) from system.numbers limit 10;
|
||
select 2 = multiMatchAnyIndex(materialize('gogleuedeuniangoogle'), ['google1', 'unian']) from system.numbers limit 10;
|
||
select 0 != multiMatchAnyIndex(materialize('gogleuedeuniangoogle'), ['.*goo.*', '.*yan.*']) from system.numbers limit 10;
|
||
select 5 = multiMatchAnyIndex(materialize('vladizlvav dabe don\'t heart me no more'), ['what', 'is', 'love', 'baby', 'no mo??', 'dont', 'h.rt me']) from system.numbers limit 10;
|
||
|
||
SELECT multiMatchAny(materialize('/odezhda-dlya-bega/'), ['/odezhda-dlya-bega/', 'kurtki-i-vetrovki-dlya-bega', 'futbolki-i-mayki-dlya-bega']);
|
||
SELECT 1 = multiMatchAny('фабрикант', ['f[ae]b[ei]rl', 'ф[иаэе]б[еэи][рпл]', 'афиукд', 'a[ft],th', '^ф[аиеэ]?б?[еэи]?$', 'берлик', 'fab', 'фа[беьв]+е?[рлко]']);
|
||
|
||
-- All indices tests
|
||
SELECT [1, 2] = arraySort(multiMatchAllIndices(materialize('gogleuedeuniangoogle'), ['.*goo.*', '.*yan.*'])) from system.numbers limit 5;
|
||
SELECT [1, 3] = arraySort(multiMatchAllIndices(materialize('gogleuedeuniangoogle'), ['.*goo.*', 'neverexisted', '.*yan.*'])) from system.numbers limit 5;
|
||
SELECT [] = multiMatchAllIndices(materialize('gogleuedeuniangoogle'), ['neverexisted', 'anotherone', 'andanotherone']) from system.numbers limit 5;
|
||
SELECT [1, 2, 3, 11] = arraySort(multiMatchAllIndices('фабрикант', ['', 'рикан', 'а', 'f[ae]b[ei]rl', 'ф[иаэе]б[еэи][рпл]', 'афиукд', 'a[ft],th', '^ф[аиеэ]?б?[еэи]?$', 'берлик', 'fab', 'фа[беьв]+е?[рлко]']));
|
||
SELECT [1] = multiMatchAllIndices(materialize('/odezhda-dlya-bega/'), ['/odezhda-dlya-bega/', 'kurtki-i-vetrovki-dlya-bega', 'futbolki-i-mayki-dlya-bega']);
|
||
SELECT [] = multiMatchAllIndices(materialize('aaaa'), ['.*aa.*aaa.*', 'aaaaaa{2}', '\(aa\){3}']);
|
||
SELECT 'All tests above must return 1, all tests below return something.';
|
||
SELECT arraySort(multiMatchAllIndices(arrayJoin(['aaaa', 'aaaaaa', 'bbbb', 'aaaaaaaaaaaaaa']), ['.*aa.*aaa.*', 'aaaaaa{2}', '(aa){3}']));
|