DuckDB ndeye SQL OLAP dhatabhesi manejimendi system mukugadzira
Kuburitswa kweiyo vhezheni itsva yeDBMS kwakaziviswa DuckDB 0.6.0, shanduro mairi data compression yakagadziridzwa, kunze kwekuti mabasa matsva akawedzerwa, pamwe nekuvandudzwa kwekuchengetedza, pakati pezvimwe zvinhu.
DuckDB batanidza SQLite zvivakwa senge compactness, kugona kubatana muchimiro cheraibhurari yakasanganiswa, kuchengetedza dhatabhesi mune imwe faira uye iri nyore CLI interface, ine maturusi uye optimizations yekuita mibvunzo yekuongorora inovhara chikamu chakakosha che data rakachengetwa, semuenzaniso, iyo inoita kuunganidzwa kwezvose zviri mukati metafura kana kubatanidza akawanda matafura makuru.
Hunhu hutsva hutsva hweDuckDB 0.6.0
Mune iyi vhezheni itsva inounzwa, inosimbiswa kuti basa rakaenderera mberi pakuvandudza chimiro chekuchengetedza, kunze kwaizvozvo dhisiki kunyora modhi yaitwa, apo kana yakakura data seti yakatakurwa mune imwe kutengeserana, iyo data inomanikidzwa uye inoyerera kune faira kubva kune dhatabhesi pasina kumirira murairo weCOMMIT kuti uite kutengeserana.
Imwe shanduko inomira pachena mushanduro itsva ndeyekuti yakawedzera rutsigiro rwekuenderana kurodha data mumatafura akasiyana, iyo inogona kuwedzera zvakanyanya kurodha kumhanya pane multicore masisitimu. Semuenzaniso, mune yekare vhezheni, kurodha dhatabhesi ine 150 miriyoni mitsara pagumi-core CPU yakatora 10 masekonzi, uye mushanduro itsva, kushanda uku kunotora masekondi gumi nemanomwe. Pane nzira mbiri dzekurodha dzakafanana: nekuchengetedza rekodhi rekodhi uye pasina kurongeka kuchengetedza.
Nekumanikidza data, iyo FSST algorithm inoshandiswa (Static Symbols Quick Table), iyo inokutendera kuti urongedze data mukati memitsara uchishandisa duramazwi rakajairika remhando yemachisi. Iko kushandiswa kweiyo algorithm nyowani inobvumidzwa kudzikisa saizi yedhatabhesi rekuyedza kubva pa761 MB kusvika 251 MB.
Kumanikidza nhamba (DOUBLE uye FLOAT) iyo Chimp uye Patas algorithms inokurudzirwa. Kuenzaniswa neyekare Gorillas algorithm, Chimp inopa yakakwira mwero wekumanikidza uye nekukurumidza decompression. Iyo Patas algorithm inosarira kumashure kweChimp maererano nekumanikidza, asi inokurumidza kukurumidza mukumhanya kwedecompression, inenge yakafanana nekuverenga data isina kudzvanywa.
Zvinotaridzawo kuti yakawedzerwa hunyanzvi hwekuyedza kurodha data kubva kuCSV mafaera mune akawanda akafanana nzizi (SET experimental_parallel_csv=true), iyo inoderedza zvakanyanya nguva yekuremerwa kwemafaira makuru eCSV. Semuenzaniso, sarudzo payakagoneswa, nguva yekurodha ye720MB CSV faira yakaderedzwa kubva pamasekonzi 3,5 kusvika masekonzi 0,6.
Yeimwe shanduko izvo zvinoratidzika kubva pane iyi nyowani vhezheni:
- Iko mukana wekuita kwakafanana kwekugadzira index uye manejimendi mashandiro akaitwa.
- SQL inopa kugona kuumba mibvunzo inotanga neizwi rekuti "KUBVA" pachinzvimbo che "SELECT". Muchiitiko ichi, mubvunzo unofungidzirwa kutanga ne "SELECT *".
- Yakawedzerwa rutsigiro rwekuti "COLUMNS" kutaura muSQL, zvichikubvumidza kuti uite oparesheni pamakoramu akawanda usingadzokorore kutaura.
- Optimized memory kushandiswa. Nekumisikidza paLinux papuratifomu, iyo jemalloc raibhurari inoshandiswa pakurangarira manejimendi. Yakanyanya kunatsiridza kuita kwehash merge mashandiro kana ndangariro dzisingakwanisi.
- Yakawedzerwa ".mode duckbox" maitiro ekubuda kuCLI, kurasa mbiru dzepakati zvichienderana nehupamhi hwemitsara yehwindo rekupedzisira). Ne ".maxrows X" parameter, unogonawo kuderedza nhamba yemitsara yekubuda.
- Iyo CLI inopa mamiriro-anoziva ekuisa otomatiki (makiyi, mazita etafura, mabasa, makoramu mazita, uye mafaira ekuisa mazita anopedzwa).
- Iyo CLI inogoneswa nekusarudzika kuratidza chiratidzo chekufambira mberi kwemubvunzo.
pakupedzisira kana uri kufarira kuziva zvakawanda nezvazvo, unogona kutarisa ruzivo Mune inotevera chinongedzo.