SQL hauv Hadoop – li cas yog nws ua hauj lwm?
Txheej txheem cej luam:
SQL rau Hadoop yog ib cov ntawv analytical cov cuab yeej uas muab lub SQL-style querying thiab xyuas cov ntaub ntawv rau tus tshiab tshaj Hadoop cov moj khaum hais lus. Tus emergence ntawm SQL rau Hadoop yog ib kev loj hlob tseem ceeb rau loj ua ntaub ntawv vim hais tias nws yuav ua rau wider pawg neeg ntse ua hauj lwm nrog lub moj khaum Hadoop ua cov ntaub ntawv ntawm khiav SQL queries rau lub enormous tagnrho cov loj tej ntaub ntawv uas Hadoop muaj dab. Obviously, tus Moj khaum Hadoop yog yav tas los tsis li ntawv puas siv tau rau neeg tshwj xeeb tshaj yog saib raws qhov peevxwm querying. Raws li txoj kev loj hlob, ntau cov cuab yeej muaj lawm los txog thiab uas promises lug txhim khu lub productivity ntawm qhauj thaum nws tawm los xyuas txog kev thiab cov ntsiab lus loj uas zoo thiab ceev. Tseem, yuav tsis muaj los sib piv cov nqi heev rau kawm lub cuab tam tias nws yog tshuaj Hmoob paub txog SQL yuav tsum ua.
Nqe lus txhais ntawm SQL nyob Hadoop
Thaum sau lawm, SQL rau Hadoop yog ib pawg ntawm daim ntaub ntawv uas tso cai rau koj khiav queries SQL-style rau ntaub ntawv loj hosted ntawm lub moj khaum Hadoop tej ntaub ntawv ua. Obviously, cov ntaub ntawv querying, retrieving thiab tsom xam kuj yooj yim rau ntawm SQL nyob Hadoop. Vim SQL yog muas thaum chiv thawj tsim kom paub database, nws yuav tsum tau muab hloov raws li cov Hadoop 1 qauv uas comprises hauv daim ntawv qhia-txo thiab Hadoop faib cov ntaub ntawv uas (HDFS) thiab cov Hadoop 2 qauv uas tsis muaj cov daim ntawv qhia-txo thiab Hadoop faib cov ntaub ntawv uas (HDFS).
Ib earliest kev muab cov SQL uas Hadoop sib txuas tau lus no ua rau cov creation ntawm warehouse Hive cov uas muaj tus HiveQL software uas txawj txhais lus SQL-style queries rau hauv MapReduce hauj lwm. Tom qab ntawd, ob daim ntaub ntawv yog tsim uas yuav ua hauj lwm zoo. Mas cov rau yam tom ntej yog laum, BigSQL, Hawq, Impala, Hadapt, Ple, H-SQL, Splice tshuab, Presto, Polybase, Txim, JethroData, Shark (Hive rau txim), thiab Tez (Hive hauv Tez).
Cas yuav SQL rau Hadoop ua hauj lwm?
SQL nyob Hadoop ua haujlwm nrog Hadoop nyob rau yam nram qab no mas:
- Connectors Hadoop puag ncig pes lus nug SQL MapReduce hom kom Hadoop to taub cov lus nug.
- Laub-down lub txim tuag SQL cov lus nug nyob rau ntawm Hadoop nyob hauv.
- Tshuab faib qhov ntim ntawm SQL queries MapReduce-HDFS tej pawg ua ke nyob rau ntawm tus workloads ntawm tus nyob ntawm cov loj loj.
No mas, tias SQL cov lus nug tsis hloov nws cov xwm; Nws yog tus Hadoop uas yog adapts rau cov lus nug ib hom ntawv nws to taub.
Kev pab sab saum toj ntawm SQL nyob Hadoop
Raws li twb tau teev, SQL ntawm Hadoop yog ib tug tseem ceeb tsim lub ntsiab lus teb kev tsom xam cov ntaub ntawv loj puas siv tau rau ntau tus neeg thiab txiav tsom xam ntaub ntawv yooj yim thiab ceev. Muaj tsis muaj tsis ntseeg tias lub moj khaum Hadoop tej ntaub ntawv nws kuj yog ib tug uas zoo rau kev loj tiam sis nws tseem tau nws tus kheej yuav tsum accessed yog ib pawg neeg xwb tsis tas vim kev loj loj ntawd ntxiv kom paub cov cim architecture tab sis tseem tau compatibility issu lawv nrog lwm tus yees. SQL hauv Hadoop promises nyob rau cov teeb meem.
Coob tus neeg nkag tau rau Hadoop tam sim no
No mas, tias SQL txog Hadoop lawm Hadoop ntau egalitarian nyob rau hauv qhov kev txiav txim zoo tias wider pawg neeg tam sim no yuav siv tau Hadoop rau ntaub ntawv thiab kev txheeb xyuas cov ntaub ntawv. Dhau, yuav kom koj siv cov Hadoop, koj yuav tsum tau kom paub txog qhov Hadoop architecture — MapReduce, Hadoop muab theej thiab faib cov ntaub ntawv kaw lus, los sis HBase. Tam sim no, koj yuav ntsaws rau yuav luag txhua analytical reporting tuam thiab kev siv thiab txheeb xyuas cov ntaub ntawv. Tsaug SQL nyob Hadoop, ib tug xov tooj ntawm SQL nyob Hadoop xyaw xws li Cloudera Impala, Txhawb lwm hom, Hadapt, CitusDB, InfiniDB, MammothDB, MemSQL, Pivotal HawQ, Laum Apache, ScleraDB, Kev kawm DataDirect, Simba thiab tshuab Splice yog tam sim no muaj muag rau cov ntaub ntawv loj. Obviously, qhov no muaj qhib Hadoop rau ib cov neeg tuaj saib wider uas muaj tam sim no cia siab tias yuav ua rau kom lawv rov qab los rau peev hauv cov ntaub ntawv loj.
Ntsiab lus loj nrog Hadoop sob tam sim no yog
Tam sim no, koj nyuam qhuav tau khiav tus zoo hluas SQL lus nug txog cov ntaub ntawv loj retrieve thiab txheeb xyuas cov ntaub ntawv. SQL muaj evolved xwb los cia li yog ib tug uas paub database rau ib tug ntawv loj tsom uas yog tseeb teeb meem. Koj tsis tas txhawj li cas Hadoop yog xyuas txog qhov queries — nws muaj nws txoj kev txhais cov SQL queries thiab muab koj cov ntsiab. Ntseeg tau tias txawm lub Hadoop faib cov ntaub ntawv kaw lus (HDFS) puas muaj rau thaum uas tig mus ua nyob hauv tej ntaub ntawv loj, nws yuav ua kom nws lub peevxwm ua yog tias nws ua haujlwm nrog SQL-style sib tham sib querying. Ua ntej lub HDFS nrog SQL, nws yuav siv sij hawm ntev mus txoj kev ntawv cov HDFS thiab cov neeg ua hauj lwm yuav tsum tau chaw me lossis cov ntaub ntawv zaum. Thiab lub queries no twb tsis yog sib tham sib. Nrog lub moj khaum Apache Tez uas comprises lub txim analytical cav thiab tus ple lus nug sib tham sib accelerator rau tus nas muv tej ntawv warehouse, tau nov lawd tej teeb meem no lawm. Raws li Anu Jain, Tus thawj tswj ntawm kev zoo thiab architecture ntawm retailer Target Corporation, "Nws tseem ceeb heev rau peb kom peb muab cov neeg siv sib tham sib tham sib zog nqus. Nrog Tez peb yeej muab uas muaj peev xwm mus rau lub lag luam."
Tej chaw uas sib tham sib analytics muaj tau cog ntawm cov neeg siv Hadoop, Raws li ib daim ntawv ntsuam xyuas Gartner ices. Raws li daim ntawv ntsuam xyuas, 32% Cov neeg teb siv kev cuam tshuam rau cov HDFS los yog HDFS los sis Hbase, 27% siv nws tus kheej tsim queries los ntawm Hive thaum 23% Siv cov cuab yeej siv tes ua noj haus xws li Cloudera Impala thiab Pivotal HAWQ.
Lwm foundective rau SQL rau Hadoop hadoop
Thaum no mas, tias SQL rau Hadoop yuav daws tau ntau yam teeb meem uas peb muaj Hadoop, Muaj lwm saib uas pom tias SQL tej zaum yuav muaj teeb meem ntau heev tshwj xeeb tshaj yog thaum lawv tshwj xeeb Hadoop. Raws li qhov pom, SQL yuav tsis tom qab txhua yam uas npaum li ib analytical cuab yeej thaum nws tawm los loj cov ntaub ntawv. Raws li Hadoop qhov ua siab tshaj cov neeg siv panelist John Hmoiams, SQL yuav tsis yog zoo analytical ua hauj lwm nrog cov ntaub ntawv loj. Raws li Hmoob, leej twg yog tus senior vice president rau platform haujlwm, Muaj tseeb tiag, uas muaj cov neeg siv lub tsheb-muas platform hauv online, "SQL execution lub sij hawm rau ib cov ntaub ntawv loj yog qeeb. Meanwhile, Hadoop-on-SQL yuav tau sai nrog tej yam zoo li Yarn thiab Tez.” Thiab uas tsis yog qhov teeb meem no tsuas muaj SQL. Muaj ntau ntau kev kawm xws li cov ntaub ntawv kawm, tswvyim, Index thiab query creation thiab normalization uas koj yuav tsum tau saib xyuas thaum koj tseem txaus SQL nrog SQL Hadoop thiab koj yuav tau siv sij hawm ntau heev thiab dag zog. Tag nrho cov uas siv zog, Tsis muaj guarante tias koj tau ua haujlwm ib tug neeg ua hauj lwm. Yog hais tias txhua yam uas tau hloov daim ntawv thov kev hloov, Tej zaum koj yuav tau thov li cas koj twb ua. Tsis siv SQL, ntaub ntawv loj-Kev loj hlob teem tau raws li Java thiab Python vim hais tias cov lus no zoo suited rau cov ntaub ntawv unstructured tej ntaub ntawv.
Txoj kev
Txiav txim tawm ntawm seb SQL rau Hadoop yog cov lus teb rau cov teeb meem neeg muag nrog siv Hadoop. Tiam sis kom meej meej, Qhov kev lag luam yuav tsum tau ib lwm txoj zoo rau hadoop tus kheej cov ntaub ntawv querying peev xwm thiab tias lwm txoj muaj kom tau sib tham sib tham. SQL rau Hadoop cov cuab yeej siv analytics uas yuav pab tau. Enterprises tsis xav mus khib nyiab mus ua kom paub nyuab, sij hawm noj analytics. Rau lub sij hawm, enterprises nrhiav SQL ntawm hadoop cov cuab yeej siv, raws li cov kev sojntsuam Gartner nyob.
