{"id":74737,"date":"2020-03-20T08:43:14","date_gmt":"2020-03-20T05:43:14","guid":{"rendered":"https:\/\/prohoster.info\/blog\/administrirovanie\/optimizacziya-strok-v-clickhouse-doklad-yandeksa"},"modified":"2020-03-20T08:43:14","modified_gmt":"2020-03-20T05:43:14","slug":"optimizacziya-strok-v-clickhouse-doklad-yandeksa","status":"publish","type":"post","link":"https:\/\/prohoster.info\/hr\/blog\/administrirovanie\/optimizacziya-strok-v-clickhouse-doklad-yandeksa","title":{"rendered":"Optimizacija nizova u ClickHouse. Izlaganje Yandexa","gt_translate_keys":[{"key":"rendered","format":"text"}]},"content":{"rendered":"<p>Analiti\u010dka SUDB ClickHouse obra\u0111uje mno\u0161tvo razli\u010ditih stringova, tro\u0161e\u0107i resurse. Kako bi se ubrzao rad sustava, stalno se dodaju nove optimizacije. Razvija\u010d ClickHousea Nikolaj Ko\u010detov govori o tipu podataka string, uklju\u010duju\u0107i novi tip, LowCardinality, i obja\u0161njava kako se mo\u017ee ubrzati rad sa stringovima. <\/p>\n<p><center><div class=\"youtube-placeholder\" data-id=\"rqf-ILRgBdY\" onclick=\"loadVideo(this)\">\r\n        <img decoding=\"async\" src=\"https:\/\/img.youtube.com\/vi\/rqf-ILRgBdY\/hqdefault.jpg\" alt=\"Pogledajte video\" loading=\"lazy\" width=\"480\" height=\"360\" style=\"width:100%;height:auto;\">\r\n        <div class=\"play-button\"><\/div>\r\n    <\/div><\/center><br \/>\n\u2014 Prvo, hajde da razjasnimo kako mo\u017eemo pohraniti stringove. <br \/>\n<noindex><a rel=\"nofollow\" name=\"habracut\"><\/a><\/noindex><br \/>\n<img decoding=\"async\" alt=\"Optimizacija nizova u ClickHouse. Izlaganje Yandexa\" src=\"\/wp-content\/uploads\/2020\/03\/5e38e903aaaf54a533bd3026e76a14be.jpg\" style=\"display:block;margin: 0 auto;\" \/><br \/>\n<br \/>\nImamo tipove podataka string. String je dobar po defaultu, treba ga koristiti skoro uvijek. Ima mali Overhead \u2014 9 bajtova po jednom stringu. Ako \u017eelimo da veli\u010dina stringova bude fiksna i unaprijed poznata, bolje je koristiti FixedString. U njemu mo\u017eemo postaviti \u017eeljeni broj bajtova, \u0161to je korisno za podatke poput IP adresa ili hash funkcija. <\/p>\n<p><img decoding=\"async\" alt=\"Optimizacija nizova u ClickHouse. Izlaganje Yandexa\" src=\"\/wp-content\/uploads\/2020\/03\/eaae898fb5163cb1c070c34a13f6f160.jpg\" style=\"display:block;margin: 0 auto;\" \/><br \/>\n<br \/>\nNaravno, ponekad ne\u0161to usporava. Recimo, pravite upit na tablici. ClickHouse \u010dita prili\u010dno veliku koli\u010dinu podataka, recimo, brzinom od 100 GB\/s, a pritom se obra\u0111uje malo stringova. Imamo dvije tablice koje sadr\u017ee gotovo jednake podatke. Iz druge tablice ClickHouse \u010dita podatke ve\u0107om brzinom, ali stringova po sekundi \u010dita tri puta manje. <\/p>\n<p><img decoding=\"async\" alt=\"Optimizacija nizova u ClickHouse. Izlaganje Yandexa\" src=\"\/wp-content\/uploads\/2020\/03\/be7cc47e628112bb212ed0e8ff9091ed.jpg\" style=\"display:block;margin: 0 auto;\" \/><br \/>\n<br \/>\nAko pogledamo veli\u010dinu komprimiranih podataka, ona \u0107e se pokazati gotovo jednakom. U stvarnosti su u tablicama zapisani isti podaci \u2014 prvi milijard brojeva \u2014 samo \u0161to su u prvom stupcu zapisani kao UInt64, a u drugom kao String. Zbog toga drugi upit du\u017ee \u010dita podatke s diska i dekomprimira ih. <\/p>\n<p><img decoding=\"async\" alt=\"Optimizacija nizova u ClickHouse. Izlaganje Yandexa\" src=\"\/wp-content\/uploads\/2020\/03\/c699103cecd972fcee73841cc35716d6.jpg\" style=\"display:block;margin: 0 auto;\" \/><br \/>\n<br \/>\nEvo jo\u0161 jednog primjera. Pretpostavimo da postoji unaprijed poznati skup stringova koji je ograni\u010den konstantom 1000 ili 10 000 i prakti\u010dki se nikada ne mijenja. Za ovaj slu\u010daj nam odgovara tip podataka Enum, u ClickHouseu ih ima dva \u2014 Enum8 i Enum16. Zbog pohrane u Enum brzo obra\u0111ujemo upite. <\/p>\n<p>U ClickHouseu postoje ubrzanja za GROUP BY, IN, DISTINCT i optimizacije za neke funkcije, na primjer za usporedbu s konstantnim stringom. Naravno, brojevi u stringu se ne pretvaraju, nego se konstantni string pretvara u vrijednost Enum. Nakon toga se sve brzo uspore\u0111uje. <\/p>\n<p>Ali postoje i nedostaci. \u010cak i ako znamo to\u010dan skup stringova, ponekad ga treba dopuniti. Kada stigne novi string \u2014 moramo napraviti ALTER. <\/p>\n<p><img decoding=\"async\" alt=\"Optimizacija nizova u ClickHouse. Izlaganje Yandexa\" src=\"\/wp-content\/uploads\/2020\/03\/064c35faaf3d9de2f12d29f83197e983.jpg\" style=\"display:block;margin: 0 auto;\" \/><br \/>\n<br \/>\nALTER za Enum u ClickHouse je maksimalno optimiziran. Ne prepisujemo podatke na disku, ali ALTER mo\u017ee usporavati zbog \u010dinjenice da se strukture Enum pohranjuju u shemu same tablice. Stoga moramo \u010dekati na zahtjeve za \u010ditanje iz tablice, na primjer. <\/p>\n<p>Postavlja se pitanje, mo\u017eemo li to u\u010diniti bolje? Vjerojatno mo\u017eemo. Mogu\u0107e je pohraniti strukturu Enum ne u shemu tablice, ve\u0107 u ZooKeeper. Me\u0111utim, mogu se pojaviti problemi s uskla\u0111ivanjem. Na primjer, jedna replika je primila podatke, druga nije, i ako ima staru verziju Enum, ne\u0161to \u0107e se pokvariti. (U ClickHouse gotovo smo zavr\u0161ili s neblokiraju\u0107im ALTER zahtjevima. Kada ih zavr\u0161imo, ne\u0107e biti potrebno \u010dekati na zahtjeve za \u010ditanje.)<\/p>\n<p><img decoding=\"async\" alt=\"Optimizacija nizova u ClickHouse. Izlaganje Yandexa\" src=\"\/wp-content\/uploads\/2020\/03\/f5d2fdf14778f6afdd4acc8c204e163c.jpg\" style=\"display:block;margin: 0 auto;\" \/><br \/>\n<br \/>\nDa bismo izbjegli muke s ALTER Enum, mo\u017eemo koristiti vanjske rje\u010dnike ClickHouse. Podsje\u0107am vas da je to struktura podataka klju\u010d-vrijednost unutar ClickHouse, koja omogu\u0107uje dohva\u0107anje podataka iz vanjskih izvora, na primjer iz MySQL tablica. <\/p>\n<p>U rje\u010dniku ClickHouse \u010duvamo mnogo razli\u010ditih nizova, a u tablici \u2014 njihove identifikatore u obliku brojeva. Ako trebamo dobiti niz, pozivamo funkciju dictGet i radimo s njom. Nakon toga ne moramo raditi ALTER. Da bismo ne\u0161to dodali u Enum, umetnemo to u istu MySQL tablicu. <\/p>\n<p>No tu se javljaju drugi problemi. Prvo, neugodna sintaksa. Ako \u017eelimo dobiti niz, moramo pozvati dictGet. Drugo, nedostatak nekih optimizacija. Usporedba s konstantnim nizom za rje\u010dnike ne ide tako brzo. <\/p>\n<p>Mogu\u0107i su i problemi s a\u017euriranjem. Pretpostavimo da smo zatra\u017eili niz u predmemoriranom rje\u010dniku, a nije stigao u predmemoriju. Tada moramo \u010dekati da se podaci u\u010ditaju iz vanjskog izvora. <\/p>\n<p><img decoding=\"async\" alt=\"Optimizacija nizova u ClickHouse. Izlaganje Yandexa\" src=\"\/wp-content\/uploads\/2020\/03\/534818ee523a243623a5cb2babe08040.jpg\" style=\"display:block;margin: 0 auto;\" \/><br \/>\n<br \/>\nZajedni\u010dki nedostatak oba metoda je to \u0161to pohranjujemo sve klju\u010deve na jednom mjestu i sinhroniziramo ih. Pa za\u0161to ne bismo pohranili rje\u010dnike lokalno? Nema sinhronizacije \u2014 nema problema. Mo\u017ee se pohraniti rje\u010dnik lokalno u dijelu diska. Drugim rije\u010dima, izvr\u0161ili smo umetanje, zapisali rje\u010dnik. Ako radimo s podacima u memoriji, mo\u017eemo zapisati rje\u010dnik ili u blok podataka, ili u komad stupca, ili u neki predmemoriju, kako bismo ubrzali izra\u010dune.<\/p>\n<h3>Rje\u010dni\u010dko kodiranje nizova <\/h3>\n<p>\nTako smo do\u0161li do stvaranja novog tipa podataka u ClickHouse \u2014 LowCardinality. To je format pohrane podataka: kako se oni zapisuju na disk i kako se \u010ditaju, kako su predstavljeni u memoriji i shema njihove obrade. <\/p>\n<p><img decoding=\"async\" alt=\"Optimizacija nizova u ClickHouse. Izlaganje Yandexa\" src=\"\/wp-content\/uploads\/2020\/03\/e25b2eb12ff158b89b7d2dbdfb8e2e5a.jpg\" style=\"display:block;margin: 0 auto;\" \/><br \/>\n<br \/>\nNa slajdu se nalaze dva stupca. S desne strane, redovi se pohranjuju standardno, kao tip String. O\u010digledno je da se radi o modelima mobilnih telefona. S lijeve strane nalazi se to\u010dno isti stupac, samo u tipu LowCardinality. Sastoji se od rje\u010dnika s mno\u0161tvom razli\u010ditih stringova (stringovi iz desnog stupca) i popisa pozicija (brojeva redova). <\/p>\n<p>Pomo\u0107u ove dvije strukture mo\u017ee se obnoviti izvorni stupac. Tako\u0111er postoji obrnuti indeks \u2014 hash-tablica koja poma\u017ee prona\u0107i poziciju u rje\u010dniku prema stringu. Ona je potrebna za ubrzanje nekih upita. Na primjer, ako \u017eelimo usporediti, potra\u017eiti string u na\u0161em stupcu ili ih spojiti. <\/p>\n<p>LowCardinality je parametarski tip podataka. Mo\u017ee biti ili broj, ili ne\u0161to \u0161to se pohranjuje kao broj, ili string, ili Nullable od njih. <\/p>\n<p><img decoding=\"async\" alt=\"Optimizacija nizova u ClickHouse. Izlaganje Yandexa\" src=\"\/wp-content\/uploads\/2020\/03\/d52ef0c99617238442d4f71e43af48af.jpg\" style=\"display:block;margin: 0 auto;\" \/><br \/>\n<br \/>\nOsobitost LowCardinality je da se mo\u017ee sa\u010duvati za neke funkcije. Na slajdu se vidi primjer upita. U prvom redu sam stvorio stupac tipa LowCardinality od String, nazvao ga S. Zatim sam ga upitao za ime \u2014 ClickHouse je rekao da je to LowCardinality od String. Sve to\u010dno.<\/p>\n<p>Tre\u0107i red je gotovo isti, samo smo pozvali funkciju length. U ClickHouseu funkcija length vra\u0107a tip podataka UInt64. Ali sada imamo LowCardinality od UInt64. Koja je svrha?<\/p>\n<p><img decoding=\"async\" alt=\"Optimizacija nizova u ClickHouse. Izlaganje Yandexa\" src=\"\/wp-content\/uploads\/2020\/03\/a4a51a0f1ecd68f7fffd521ab786c8a5.jpg\" style=\"display:block;margin: 0 auto;\" \/><br \/>\n<br \/>\nU rje\u010dniku su se nalazila imena mobilnih telefona, primijenili smo funkciju length. Sada imamo sli\u010dan rje\u010dnik, sastavljen samo od brojeva \u2014 to su du\u017eine stringova. Stupac s pozicijama nije se promijenio. Na kraju smo obradili manje podataka, u\u0161tedjeli na vremenu upita. <\/p>\n<p>Mogu biti i druge optimizacije, na primjer dodavanje jednostavnog ke\u0161a. Pri izra\u010dunavanju vrijednosti funkcije mo\u017eemo je zapamtiti i sastaviti je ponovo, ne izra\u010dunavaju\u0107i je ponovno. <\/p>\n<p>Tako\u0111er mo\u017ee biti napravljena optimizacija GROUP BY, jer je na\u0161 stupac s rje\u010dnikom ve\u0107 djelomi\u010dno agregiran \u2014 mo\u017eete br\u017ee izra\u010dunati vrijednosti hash-funkcija i otprilike prona\u0107i bucket u koji smjestiti sljede\u0107i red. Tako\u0111er se mogu specializirati neke agregatne funkcije, na primjer uniq, jer se u nju mo\u017ee poslati samo rje\u010dnik, dok pozicije ostaju netaknute \u2014 tako \u0107e sve raditi br\u017ee. Prve dvije optimizacije smo ve\u0107 dodali u ClickHouse.<\/p>\n<p><img decoding=\"async\" alt=\"Optimizacija nizova u ClickHouse. Izlaganje Yandexa\" src=\"\/wp-content\/uploads\/2020\/03\/943e0a8ae5a37cd47164246db3a67893.jpg\" style=\"display:block;margin: 0 auto;\" \/><br \/>\n<br \/>\n\u0160to ako stvorimo kolonu s na\u0161im vrstama podataka i umetnemo u nju mnogo lo\u0161ih razli\u010ditih redaka? Ho\u0107e li nam se memorija prepuniti? Ne, za to ClickHouse ima dva posebna pode\u0161avanja. Prvo je low_cardinality_max_dictionary_size. To je maksimalna veli\u010dina rje\u010dnika koja se mo\u017ee zapisati na disk. Umetanje se odvija na sljede\u0107i na\u010din: kada umetnemo podatke, dolazi nam tok redaka, od kojih oblikujemo veliki zajedni\u010dki rje\u010dnik. Ako rje\u010dnik postane ve\u0107i od vrijednosti ovog pode\u0161avanja, zapisujemo trenutni rje\u010dnik na disk, a ostale retke pohranjujemo negdje <\/p>\n<p>drugo <\/p>\n<p>Kada se prepuniti, ponovno \u0107emo ga zapisati na disk i po\u010deti stvarati tre\u0107i. Ovo pode\u0161avanje isklju\u010duje tu mogu\u0107nost prema zadanim postavkama. <\/p>\n<p>U stvari, mnogo rje\u010dnika mo\u017ee biti korisno ako \u017eelimo umetnuti neku skupinu redaka, ali slu\u010dajno umetnemo \"sme\u0107e\". Recimo, prvo smo umetnuli lo\u0161e redove, a zatim dobre. Tada \u0107e se rje\u010dnik podijeliti na mnogo malih rje\u010dnika. Neki od njih \u0107e sadr\u017eavati \"sme\u0107e\", ali posljednji \u0107e biti s dobrim redovima. I ako budemo \u010ditali, recimo, samo posljednju granulaciju, sve \u0107e opet raditi brzo. <\/p>\n<p><img decoding=\"async\" alt=\"Optimizacija nizova u ClickHouse. Izlaganje Yandexa\" src=\"\/wp-content\/uploads\/2020\/03\/834ba3295ace98870334aa10a78e4781.jpg\" style=\"display:block;margin: 0 auto;\" \/><br \/>\n<br \/>\nPrije nego \u0161to govorimo o prednostima LowCardinality, odmah \u0107u re\u0107i da vjerojatno ne\u0107emo posti\u0107i smanjenje podataka na disku (iako se to mo\u017ee dogoditi), jer ClickHouse komprimira podatke. Postoji zadani na\u010din - LZ4. Tako\u0111er se mo\u017ee provesti kompresija pomo\u0107u ZSTD. No, oba algoritma ve\u0107 implementiraju rje\u010dnu kompresiju, pa na\u0161 vanjski rje\u010dnik ClickHouse ne\u0107e pomo\u0107i previ\u0161e. <\/p>\n<p>Da ne budem neosnovan, uzeo sam neke podatke iz metrike \u2014 String, LowCardinality(String) i Enum \u2014 i sa\u010duvao ih u razli\u010ditim tipovima podataka. Dobio sam tri kolone, u kojima je zapisano milijardu redaka. U prvoj koloni, CodePage, ukupno 62 vrijednosti. Vidimo da nam je kod LowCardinality(String) bolja kompresija. String je ne\u0161to lo\u0161iji, ali to je vjerojatno zbog toga \u0161to su nizovi kratki, \u010duvamo njihove duljine, a one zauzimaju puno mjesta i slabo se komprimiraju. <\/p>\n<p>Ako uzmemo PhoneModel, ima ih 48 tisu\u0107a \u2014 ve\u0107 vi\u0161e, i razlike izme\u0111u String i LowCardinality(String) gotovo da nema. Za URL smo tako\u0111er u\u0161tedjeli samo 2 GB \u2014 mislim da na to ne vrijedi ra\u010dunati. <\/p>\n<h3>Procjena brzine rada<\/h3>\n<p>\n<img decoding=\"async\" alt=\"Optimizacija nizova u ClickHouse. Izlaganje Yandexa\" src=\"\/wp-content\/uploads\/2020\/03\/0b646570639f8f2e29b599473b84e25c.jpg\" style=\"display:block;margin: 0 auto;\" \/><br \/>\n<b><noindex><a rel=\"nofollow\" href=\"https:\/\/github.com\/toddwschneider\/nyc-taxi-data\">Veza sa slajda<\/a><\/noindex><\/b><\/p>\n<p>Sada \u0107emo procijeniti brzinu rada. Da bih to procijenio, koristio sam skup podataka s opisom vo\u017enji taksijem u New Yorku. On <noindex><a rel=\"nofollow\" href=\"https:\/\/github.com\/toddwschneider\/nyc-taxi-data\">dostupan<\/a><\/noindex> se nalazi na GitHubu. U njemu se nalazi ne\u0161to vi\u0161e od milijarde vo\u017enji. Osim toga, tu su prikazane lokacija, vrijeme po\u010detka i zavr\u0161etka vo\u017enje, na\u010din pla\u0107anja, broj putnika i \u010dak tip taksija \u2014 zeleni, \u017euti i Uber. <\/p>\n<p><img decoding=\"async\" alt=\"Optimizacija nizova u ClickHouse. Izlaganje Yandexa\" src=\"\/wp-content\/uploads\/2020\/03\/0901669dec01b0f956d6976e3aa8b741.jpg\" style=\"display:block;margin: 0 auto;\" \/><br \/>\n<br \/>\nPrvi upit sam napravio prili\u010dno jednostavno \u2014 pitao sam gdje se naj\u010de\u0161\u0107e naru\u010duju taksiji. Za to treba uzeti lokaciju s koje je naru\u010deno, napraviti GROUP BY i prebrojati funkciju count. ClickHouse ne\u0161to daje. <\/p>\n<p><img decoding=\"async\" alt=\"Optimizacija nizova u ClickHouse. Izlaganje Yandexa\" src=\"\/wp-content\/uploads\/2020\/03\/50983341e20bcbc1b5adbaed35f8624d.jpg\" style=\"display:block;margin: 0 auto;\" \/><br \/>\n<br \/>\nDa bih izmjerio brzinu obrade upita, stvorio sam tri tablice s istim podacima, ali sam za na\u0161u po\u010detnu lokaciju koristio tri razli\u010dita tipa podataka \u2014 String, LowCardinality i Enum. LowCardinality i Enum su se pokazali pet puta br\u017eima od Stringa. Enum je br\u017ei jer radi s brojevima, a LowCardinality zbog optimizacije GROUP BY. <\/p>\n<p><img decoding=\"async\" alt=\"Optimizacija nizova u ClickHouse. Izlaganje Yandexa\" src=\"\/wp-content\/uploads\/2020\/03\/e2ce4f2127d8728287488c85c92085fc.jpg\" style=\"display:block;margin: 0 auto;\" \/><br \/>\n<br \/>\nPojdimo u slo\u017eeniji upit \u2014 pitamo se gdje se nalazi najpopularniji park u New Yorku. Opet \u0107emo to mjeriti prema tome gdje se naj\u010de\u0161\u0107e naru\u010duju taksiji, ali \u0107emo filtrirati samo one lokacije koje sadr\u017ee rije\u010d 'park'. Tako\u0111er \u0107emo dodati funkciju like. <\/p>\n<p><img decoding=\"async\" alt=\"Optimizacija nizova u ClickHouse. Izlaganje Yandexa\" src=\"\/wp-content\/uploads\/2020\/03\/7cf6fd64751f3ae578581caddde17aaf.jpg\" style=\"display:block;margin: 0 auto;\" \/><br \/>\n<br \/>\nGledamo vrijeme \u2014 vidimo da Enum iznenada po\u010dinje usporavati. \u010cak radi sporije od standardnog tipa podataka String. To se doga\u0111a zato \u0161to funkcija like nije optimizirana za Enum. Moramo pretvoriti na\u0161e nizove iz Enum u obi\u010dne nizove \u2014 radimo vi\u0161e posla. LowCardinality(String) tako\u0111er nije optimiziran prema zadanim postavkama, ali tu like radi nad rje\u010dnikom, zbog \u010dega se upit ubrzava u odnosu na String. <\/p>\n<p>Kada radimo s Enum-om, postoji ve\u0107i problem. Ako ga \u017eelimo optimizirati, to moramo u\u010diniti svugdje u kodu. Pretpostavimo da smo napisali novu funkciju \u2014 nu\u017eno je smisliti optimizaciju za Enum. Dok je u LowCardinality sve optimizirano prema zadanim postavkama.<\/p>\n<p><img decoding=\"async\" alt=\"Optimizacija nizova u ClickHouse. Izlaganje Yandexa\" src=\"\/wp-content\/uploads\/2020\/03\/459243eb5b3dd993393d0184b224ea09.jpg\" style=\"display:block;margin: 0 auto;\" \/><br \/>\n<br \/>\nPogledajmo posljednji upit, koji je vi\u0161e umjetan. Jednostavno \u0107emo izra\u010dunati hash funkciju za na\u0161u lokaciju. Hash funkcija je prili\u010dno spor upit, izra\u010dunava se dugo, pa \u0107e sve usporiti tri puta.<\/p>\n<p><img decoding=\"async\" alt=\"Optimizacija nizova u ClickHouse. Izlaganje Yandexa\" src=\"\/wp-content\/uploads\/2020\/03\/a5f49c22748971646ec358eb52d3feda.jpg\" style=\"display:block;margin: 0 auto;\" \/><br \/>\n<br \/>\nLowCardinality jo\u0161 uvijek radi br\u017ee, iako ovdje nema filtriranja. To se doga\u0111a zato \u0161to na\u0161e funkcije rade samo s rje\u010dnikom. Funkcija za izra\u010dun hash-a ima jedan argument \u2014 mo\u017ee obraditi manje podataka i tako\u0111er mo\u017ee vratiti LowCardinality. <\/p>\n<p><img decoding=\"async\" alt=\"Optimizacija nizova u ClickHouse. Izlaganje Yandexa\" src=\"\/wp-content\/uploads\/2020\/03\/7842b7eb2debb4a4b1d4cbc3488deb32.jpg\" style=\"display:block;margin: 0 auto;\" \/><br \/>\n<br \/>\nNa\u0161 globalni cilj je posti\u0107i brzinu rada ne manju od one kod String-a u svim slu\u010dajevima, i zadr\u017eati ubranja. Mo\u017eda \u0107emo jednog dana zamijeniti String s LowCardinality, a\u017eurirat \u0107ete ClickHouse i sve \u0107e raditi malo br\u017ee.<br \/>\n<br \/>Izvor: <a content=\"nofollow\" rel=\"nofollow\" href=\"https:\/\/habr.com\/ru\/company\/yandex\/blog\/492868\/\">habr.com<\/a> <\/p>","protected":false,"gt_translate_keys":[{"key":"rendered","format":"html"}]},"excerpt":{"rendered":"<p>\u0410\u043d\u0430\u043b\u0438\u0442\u0438\u0447\u0435\u0441\u043a\u0430\u044f \u0421\u0423\u0411\u0414 ClickHouse \u043e\u0431\u0440\u0430\u0431\u0430\u0442\u044b\u0432\u0430\u0435\u0442 \u043c\u043d\u043e\u0436\u0435\u0441\u0442\u0432\u043e \u0440\u0430\u0437\u043d\u044b\u0445 \u0441\u0442\u0440\u043e\u043a, \u043f\u043e\u0442\u0440\u0435\u0431\u043b\u044f\u044f \u0440\u0435\u0441\u0443\u0440\u0441\u044b. \u0414\u043b\u044f \u0443\u0441\u043a\u043e\u0440\u0435\u043d\u0438\u044f \u0440\u0430\u0431\u043e\u0442\u044b \u0441\u0438\u0441\u0442\u0435\u043c\u044b \u043f\u043e\u0441\u0442\u043e\u044f\u043d\u043d\u043e \u0434\u043e\u0431\u0430\u0432\u043b\u044f\u044e\u0442\u0441\u044f \u043d\u043e\u0432\u044b\u0435 \u043e\u043f\u0442\u0438\u043c\u0438\u0437\u0430\u0446\u0438\u0438. \u0420\u0430\u0437\u0440\u0430\u0431\u043e\u0442\u0447\u0438\u043a ClickHouse \u041d\u0438\u043a\u043e\u043b\u0430\u0439 \u041a\u043e\u0447\u0435\u0442\u043e\u0432 \u0440\u0430\u0441\u0441\u043a\u0430\u0437\u044b\u0432\u0430\u0435\u0442 \u043e \u0441\u0442\u0440\u043e\u043a\u043e\u0432\u043e\u043c \u0442\u0438\u043f\u0435 \u0434\u0430\u043d\u043d\u044b\u0445, \u0432 \u0442\u043e\u043c \u0447\u0438\u0441\u043b\u0435 \u043e \u043d\u043e\u0432\u043e\u043c \u0442\u0438\u043f\u0435, LowCardinality, \u0438 \u043e\u0431\u044a\u044f\u0441\u043d\u044f\u0435\u0442, \u043a\u0430\u043a \u043c\u043e\u0436\u043d\u043e \u0443\u0441\u043a\u043e\u0440\u0438\u0442\u044c \u0440\u0430\u0431\u043e\u0442\u0443 \u0441\u043e \u0441\u0442\u0440\u043e\u043a\u0430\u043c\u0438. \u2014 \u0421\u043d\u0430\u0447\u0430\u043b\u0430 \u0434\u0430\u0432\u0430\u0439\u0442\u0435 \u0440\u0430\u0437\u0431\u0435\u0440\u0435\u043c\u0441\u044f, \u043a\u0430\u043a \u043c\u043e\u0436\u043d\u043e \u0445\u0440\u0430\u043d\u0438\u0442\u044c \u0441\u0442\u0440\u043e\u043a\u0438. \u0423 \u043d\u0430\u0441 \u0435\u0441\u0442\u044c \u0441\u0442\u0440\u043e\u043a\u043e\u0432\u044b\u0435 \u0442\u0438\u043f\u044b \u0434\u0430\u043d\u043d\u044b\u0445. [&hellip;]<\/p>\n","protected":false,"gt_translate_keys":[{"key":"rendered","format":"html"}]},"author":1,"featured_media":74738,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[688],"tags":[],"class_list":["post-74737","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-administrirovanie"],"aioseo_notices":[],"aioseo_head":"\n\t\t<!-- All in One SEO 5.0.5 - aioseo.com -->\n\t<meta name=\"description\" content=\"\u0410\u043d\u0430\u043b\u0438\u0442\u0438\u0447\u0435\u0441\u043a\u0430\u044f \u0421\u0423\u0411\u0414 ClickHouse \u043e\u0431\u0440\u0430\u0431\u0430\u0442\u044b\u0432\u0430\u0435\u0442 \u043c\u043d\u043e\u0436\u0435\u0441\u0442\u0432\u043e \u0440\u0430\u0437\u043d\u044b\u0445 \u0441\u0442\u0440\u043e\u043a, \u043f\u043e\u0442\u0440\u0435\u0431\u043b\u044f\u044f \u0440\u0435\u0441\u0443\u0440\u0441\u044b. \u0414\u043b\u044f \u0443\u0441\u043a\u043e\u0440\u0435\u043d\u0438\u044f \u0440\u0430\u0431\u043e\u0442\u044b \u0441\u0438\u0441\u0442\u0435\u043c\u044b \u043f\u043e\u0441\u0442\u043e\u044f\u043d\u043d\u043e \u0434\u043e\u0431\u0430\u0432\u043b\u044f\u044e\u0442\u0441\u044f \u043d\u043e\u0432\u044b\u0435 \u043e\u043f\u0442\u0438\u043c\u0438\u0437\u0430\u0446\u0438\u0438.\" \/>\n\t<meta name=\"robots\" content=\"max-image-preview:large\" \/>\n\t<meta name=\"author\" content=\"Yuri Gagarin\"\/>\n\t<link rel=\"canonical\" href=\"https:\/\/prohoster.info\/hr\/blog\/administrirovanie\/optimizacziya-strok-v-clickhouse-doklad-yandeksa\" \/>\n\t\t<meta name=\"generator\" content=\"All in One SEO (AIOSEO) 5.0.5\" \/>\n\t\t<meta property=\"og:locale\" content=\"hr_HR\" \/>\n\t\t<meta property=\"og:site_name\" content=\"ProHoster | \u041a\u0443\u043f\u0438\u0442\u044c \u043d\u0430\u0434\u0435\u0436\u043d\u044b\u0439 \u0445\u043e\u0441\u0442\u0438\u043d\u0433 \u0434\u043b\u044f \u0441\u0430\u0439\u0442\u043e\u0432 \u0441 \u0437\u0430\u0449\u0438\u0442\u043e\u0439 \u043e\u0442 DDoS, VPS VDS \u0441\u0435\u0440\u0432\u0435\u0440\u044b\" \/>\n\t\t<meta property=\"og:type\" content=\"article\" \/>\n\t\t<meta property=\"og:title\" content=\"\ud83e\udd47\u041e\u043f\u0442\u0438\u043c\u0438\u0437\u0430\u0446\u0438\u044f \u0441\u0442\u0440\u043e\u043a \u0432 ClickHouse. \u0414\u043e\u043a\u043b\u0430\u0434 \u042f\u043d\u0434\u0435\u043a\u0441\u0430 | ProHoster\" \/>\n\t\t<meta property=\"og:description\" content=\"\u0410\u043d\u0430\u043b\u0438\u0442\u0438\u0447\u0435\u0441\u043a\u0430\u044f \u0421\u0423\u0411\u0414 ClickHouse \u043e\u0431\u0440\u0430\u0431\u0430\u0442\u044b\u0432\u0430\u0435\u0442 \u043c\u043d\u043e\u0436\u0435\u0441\u0442\u0432\u043e \u0440\u0430\u0437\u043d\u044b\u0445 \u0441\u0442\u0440\u043e\u043a, \u043f\u043e\u0442\u0440\u0435\u0431\u043b\u044f\u044f \u0440\u0435\u0441\u0443\u0440\u0441\u044b. \u0414\u043b\u044f \u0443\u0441\u043a\u043e\u0440\u0435\u043d\u0438\u044f \u0440\u0430\u0431\u043e\u0442\u044b \u0441\u0438\u0441\u0442\u0435\u043c\u044b \u043f\u043e\u0441\u0442\u043e\u044f\u043d\u043d\u043e \u0434\u043e\u0431\u0430\u0432\u043b\u044f\u044e\u0442\u0441\u044f \u043d\u043e\u0432\u044b\u0435 \u043e\u043f\u0442\u0438\u043c\u0438\u0437\u0430\u0446\u0438\u0438.\" \/>\n\t\t<meta property=\"og:url\" content=\"https:\/\/prohoster.info\/hr\/blog\/administrirovanie\/optimizacziya-strok-v-clickhouse-doklad-yandeksa\" \/>\n\t\t<meta property=\"og:image\" content=\"https:\/\/prohoster.info\/wp-content\/uploads\/2021\/11\/logo-350.jpg\" \/>\n\t\t<meta property=\"og:image:secure_url\" content=\"https:\/\/prohoster.info\/wp-content\/uploads\/2021\/11\/logo-350.jpg\" \/>\n\t\t<meta property=\"og:image:width\" content=\"350\" \/>\n\t\t<meta property=\"og:image:height\" content=\"350\" \/>\n\t\t<meta property=\"article:published_time\" content=\"2020-03-20T05:43:14+00:00\" \/>\n\t\t<meta property=\"article:modified_time\" content=\"2020-03-20T05:43:14+00:00\" \/>\n\t\t<meta property=\"article:publisher\" content=\"https:\/\/www.facebook.com\/prohoster\" \/>\n\t\t<meta property=\"article:author\" content=\"https:\/\/www.facebook.com\/prohoster\" \/>\n\t\t<!-- All in One SEO -->\n\n","aioseo_head_json":{"title":"\ud83e\udd47Optimizacija stringova u ClickHouse-u. Izvje\u0161taj Yandexa | ProHoster","description":"Analiti\u010dka SGBD ClickHouse obra\u0111uje veliki broj razli\u010ditih stringova, Tro\u0161e\u0107i resurse. Za ubrzanje sustava stalno se dodaju nove optimizacije.","canonical_url":"https:\/\/prohoster.info\/hr\/blog\/administrirovanie\/optimizacziya-strok-v-clickhouse-doklad-yandeksa","robots":"max-image-preview:large","keywords":"","webmasterTools":{"miscellaneous":""},"schema":null,"og:locale":"hr_HR","og:site_name":"ProHoster | \u041a\u0443\u043f\u0438\u0442\u044c \u043d\u0430\u0434\u0435\u0436\u043d\u044b\u0439 \u0445\u043e\u0441\u0442\u0438\u043d\u0433 \u0434\u043b\u044f \u0441\u0430\u0439\u0442\u043e\u0432 \u0441 \u0437\u0430\u0449\u0438\u0442\u043e\u0439 \u043e\u0442 DDoS, VPS VDS \u0441\u0435\u0440\u0432\u0435\u0440\u044b","og:type":"article","og:title":"\ud83e\udd47\u041e\u043f\u0442\u0438\u043c\u0438\u0437\u0430\u0446\u0438\u044f \u0441\u0442\u0440\u043e\u043a \u0432 ClickHouse. \u0414\u043e\u043a\u043b\u0430\u0434 \u042f\u043d\u0434\u0435\u043a\u0441\u0430 | ProHoster","og:description":"\u0410\u043d\u0430\u043b\u0438\u0442\u0438\u0447\u0435\u0441\u043a\u0430\u044f \u0421\u0423\u0411\u0414 ClickHouse \u043e\u0431\u0440\u0430\u0431\u0430\u0442\u044b\u0432\u0430\u0435\u0442 \u043c\u043d\u043e\u0436\u0435\u0441\u0442\u0432\u043e \u0440\u0430\u0437\u043d\u044b\u0445 \u0441\u0442\u0440\u043e\u043a, \u043f\u043e\u0442\u0440\u0435\u0431\u043b\u044f\u044f \u0440\u0435\u0441\u0443\u0440\u0441\u044b. \u0414\u043b\u044f \u0443\u0441\u043a\u043e\u0440\u0435\u043d\u0438\u044f \u0440\u0430\u0431\u043e\u0442\u044b \u0441\u0438\u0441\u0442\u0435\u043c\u044b \u043f\u043e\u0441\u0442\u043e\u044f\u043d\u043d\u043e \u0434\u043e\u0431\u0430\u0432\u043b\u044f\u044e\u0442\u0441\u044f \u043d\u043e\u0432\u044b\u0435 \u043e\u043f\u0442\u0438\u043c\u0438\u0437\u0430\u0446\u0438\u0438.","og:url":"https:\/\/prohoster.info\/hr\/blog\/administrirovanie\/optimizacziya-strok-v-clickhouse-doklad-yandeksa","og:image":"https:\/\/prohoster.info\/wp-content\/uploads\/2021\/11\/logo-350.jpg","og:image:secure_url":"https:\/\/prohoster.info\/wp-content\/uploads\/2021\/11\/logo-350.jpg","og:image:width":350,"og:image:height":350,"article:published_time":"2020-03-20T05:43:14+00:00","article:modified_time":"2020-03-20T05:43:14+00:00","article:publisher":"https:\/\/www.facebook.com\/prohoster","article:author":"https:\/\/www.facebook.com\/prohoster"},"aioseo_meta_data":{"post_id":"74737","title":null,"description":null,"keywords":null,"keyphrases":null,"primary_term":null,"canonical_url":null,"og_title":null,"og_description":null,"og_object_type":"default","og_image_type":"default","og_image_url":null,"og_image_width":null,"og_image_height":null,"og_image_custom_url":null,"og_image_custom_fields":null,"og_video":null,"og_custom_url":null,"og_article_section":null,"og_article_tags":null,"twitter_use_og":false,"twitter_card":"default","twitter_image_type":"default","twitter_image_url":null,"twitter_image_custom_url":null,"twitter_image_custom_fields":null,"twitter_title":null,"twitter_description":null,"schema":{"blockGraphs":[],"customGraphs":[],"default":{"data":{"Article":[],"Course":[],"Dataset":[],"FAQPage":[],"Movie":[],"Person":[],"Product":[],"ProductReview":[],"Car":[],"Recipe":[],"Service":[],"SoftwareApplication":[],"WebPage":[]},"graphName":"","isEnabled":true},"graphs":[]},"schema_type":null,"schema_type_options":null,"pillar_content":false,"robots_default":true,"robots_noindex":false,"robots_noarchive":false,"robots_nosnippet":false,"robots_nofollow":false,"robots_noimageindex":false,"robots_noodp":false,"robots_notranslate":false,"robots_max_snippet":null,"robots_max_videopreview":null,"robots_max_imagepreview":"large","priority":null,"frequency":null,"local_seo":null,"seo_analyzer_scan_date":null,"breadcrumb_settings":null,"limit_modified_date":false,"reviewed_by":null,"ai":null,"created":"2021-02-28 17:23:10","updated":"2022-09-27 14:48:44","focus_keyword":null,"additional_keywords":null,"truseo_locale":null},"gt_translate_keys":[{"key":"link","format":"url"}],"_links":{"self":[{"href":"https:\/\/prohoster.info\/hr\/wp-json\/wp\/v2\/posts\/74737","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/prohoster.info\/hr\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/prohoster.info\/hr\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/prohoster.info\/hr\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/prohoster.info\/hr\/wp-json\/wp\/v2\/comments?post=74737"}],"version-history":[{"count":0,"href":"https:\/\/prohoster.info\/hr\/wp-json\/wp\/v2\/posts\/74737\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/prohoster.info\/hr\/wp-json\/wp\/v2\/media\/74738"}],"wp:attachment":[{"href":"https:\/\/prohoster.info\/hr\/wp-json\/wp\/v2\/media?parent=74737"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/prohoster.info\/hr\/wp-json\/wp\/v2\/categories?post=74737"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/prohoster.info\/hr\/wp-json\/wp\/v2\/tags?post=74737"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}