ClickHouse - inooneka nekukurumidza uye intuitive data kuongororwa muTabix. Igor Stryhar

Ini ndinokurudzira kuti uverenge zvinyorwa zve2017 mushumo naIgor Stryhar "ClickHouse - inooneka nekukurumidza uye yakajeka data data muTabix."

Webhu interface yeClickHouse muTabix chirongwa.
Zvinhu zvakakosha:

• Inoshanda neClickHouse zvakananga kubva kubrowser, pasina chikonzero chekuisa imwe software;
• Mubvunzo mupepeti ane syntax inosimudzira;
• Autocompletion yemirairo;
• Zvishandiso zvekuongororwa kwegraphical kwekuita mubvunzo;
• Zvirongwa zvemavara zvekusarudza kubva.
  

Ini ndiri technical director we SMI2. Isu tiri muunganidzwa wenhau wenhau aggregator. Isu tinochengeta data rakawanda ratinogashira kubva kune vatinoshanda navo uye nekurinyoresa muClickHouse - zvingangoita zviuru makumi matatu zvikumbiro pasekondi.

Aya ndiwo data akadai se:

• Dzvanya panhau.
• Nhau dzinoratidzwa muaggregator.
• Banner inoratidzwa panetiweki yedu.
• Uye isu tinonyoresa zviitiko kubva pakaunda yedu, iyo yakafanana neYandex.Metrica. Iyi ndiyo yedu pachedu micro-analytics.

Taive nehupenyu hwekubatikana pamberi peClickHouse. Takatambura zvakanyanya, tichiedza kuchengeta iyi data pane imwe nzvimbo uye neimwe nzira kuiongorora.

Hupenyu pamberi peClickHouse - infiniDB

Chinhu chekutanga chataive nacho yaive infiniDB. Akagara nesu kwemakore mana. Takaivhura zvakaoma.

• Iyo haitsigire kubatanidza kana kupatsanura. Hapana zvinhu zvakangwara zvakadaro zvakabuda mubhokisi nekukasira.
• Ane dambudziko kurodha data. Chete chaiyo console yekushandisa iyo inogona kungoisa CSV mafaera uye chete neimwe nzira isinganyatso kujeka.
• Iyo database ndeye-single-threaded. Unogona kunyora kana kuverenga. Asi zvakaita kuti zvikwanise kugadzirisa huwandu hukuru hwe data.
• Uye aivewo netsvimbo inofadza. Husiku hwega hwega sevha yaifanira kutangwazve, zvikasadaro yaisashanda.

Akatishandira kusvika pakupera kwa2016, patakachinja zvachose kuenda kuClickHouse.

Hupenyu pamberi peClickHouse - Cassandra

Sezvo infiniDB yaive imwe-tambo, isu takasarudza kuti taida imwe mhando yedhatabheti yakawanda-yakawanda umo taigona kunyora tambo dzakawanda panguva imwe chete.

Takaedza zvinhu zvakawanda zvinofadza. Ipapo takasarudza kuedza Cassandra. Zvese zvaive zvakanaka naCassandra. 10 zvikumbiro pasekondi yega yega. 000 zvikumbiro kune imwe nzvimbo yekuverenga.

Asi aivewo nezvaaidawo. Kamwe chete pamwedzi kana kamwe chete mwedzi miviri yega yega aiona dhatabhesi desynchronization. And I had to wake up and run to fix Cassandra. Masevha akatangazve rimwe nerimwe. Uye zvinhu zvose zvakave zvakatsetseka uye zvakanaka.

Hupenyu pamberi peClickHouse - Druid

Ipapo takaona kuti taifanira kunyora mamwe data. Muna 2016 takatanga kuona Druid.

Druid ndeye yakavhurika sosi software yakanyorwa muJava. Zvakananga chaizvo. Uye yaive yakakodzera Clickstream, patinenge tichida kuchengetedza imwe mhando yezviitiko uye tozoita aggregation pazviri kana kuita analytical mishumo.

Druid aive neshanduro 0.9.X.

Iyo database pachayo yakaoma kwazvo kuendesa. Uku ndiko kuoma kwezvivakwa. Kuti uiise, zvaive zvakakodzera kuisa yakawanda, yakawanda yesimbi. Uye chidimbu chega chega chehardware chaive nemhosva yebasa rayo rakasiyana.

Kuisa data mairi, zvaive zvakakodzera kushandisa imwe mhando ye shamanism. Pane chirongwa cheOpenSource - Kugadzikana, uko kwaive kurasikirwa nedata kubva kwatiri murukova. Patakaisa data mariri, yakarasika.

Asi neimwe nzira takatanga kuishandisa. Isu, sehedgehogs vakatora zvinodhaka asi vakaramba vachidya cactus, takatanga kuizivisa. Zvakatitorera unenge mwedzi kuti tigadzirire zvivakwa zvacho. Kureva, kurongeka maseva, gadzirisa mabasa, uye zvizere otomatiki kutumira. Ndiko kuti, kana pakakundikana sumbu, sumbu rechipiri rinozoiswa otomatiki.

Asi chishamiso chakaitika. Ndakanga ndiri pazororo uye vandinoshanda navo vakanditumira link habr, iyo inotaura kuti Yandex yakasarudza kuvhura ClickHouse. Ndinoti ngatiedzei.

Uye chaizvo mumazuva maviri takatumira ClickHouse test cluster. Takatanga kurodha data imomo. Kuenzaniswa ne infiniDB, ichi chekutanga; kana ichienzaniswa neDruid, ichi chekutanga. Kuenzaniswa naCassandra, zvakare ndeyekutanga. Nekuti kana iwe ukarodha data kubva kuPP kuenda kuCassandra, saka ichi hachisi chekutanga.

Takawanei? Kuita mukumhanya. Kuita mukuchengetedza data. Ndiko kuti, nzvimbo shoma yedhisiki inoshandiswa. ClickHouse inokurumidza, inokurumidza zvikuru kana ichienzaniswa nezvimwe zvigadzirwa.

Panguva yekutanga, Yandex payakaburitsa ClickHouse muOpenSource, kwaingova nemutengi wekoni. Isu pakambani yedu SMI2 takasarudza kuedza kuita mutengi wekuzvarwa kwewebhu, kuitira kuti tikwanise kuvhura peji kubva kune browser, kunyora chikumbiro uye kuwana mhedzisiro, nekuti takatanga kunyora zvakawanda zvikumbiro. Kunyora mu console kwakaoma. Uye isu takagadzira yedu yekutanga vhezheni.

Uye pane imwe nzvimbo padyo nechando chegore rapfuura, zvechitatu-bato maturusi ekushanda neClickHouse akatanga kuoneka. Aya ndiwo maturusi akadai se:

• Redash.IO.
• Apache Zeppelin.
• Superset Airbnb. (Zvino https://github.com/apache/incubator-superset)
• Jetbrains IDE.
• Vertamedia/clickhouse-grafana.

Ndichatarisa zvimwe zvezvishandiso izvi, kureva, izvo zvandakashanda nazvo.

Chishandiso chakanaka, asi cheDruid. Paiitwa Druid, ini ndaiyedza SuperSet. Ndaimufarira. KuDruid inokurumidza chaizvo.

Izvo hazvina kukodzera ClickHouse. Ndiko kuti, inokodzera, inotanga, asi yakagadzirira kugadzirisa chete mibvunzo yekutanga senge: SARUDZA chiitiko, GROUP BY chiitiko. Iyo haitsigire iyo yakanyanya kuoma ClickHouse syntax.

Chishandiso chinotevera ndiApache Zeppelin. Ichi chinhu chakanaka uye chinonakidza. Mabasa. Inotsigira zvinyorwa, dashboards, uye inotsigira zvakasiyana-siyana. Ndinoziva mumwe munhu munharaunda yeClickHouse anoishandisa.

Asi hapana tsigiro yeClickHouse syntax, kureva kuti uchafanirwa kunyora mibvunzo mungave mukoni kana kumwewo. Tevere, tarisa kuti zvese zvinoshanda. Zvinongonetsa. Asi ine yakanaka dashboard rutsigiro.

Chishandiso chinotevera Redash.IO. Redash inobatwa paInternet. Ndiko kuti, kusiyana nemidziyo yapfuura, haidi kuiswa. Uye iyi idhibhodhi ine kugona kubatanidza data kubva kwakasiyana DataSource. Ndiko kuti, unogona kudhawunirodha kubva kuClickHouse, kubva kuMySQL, kubva kuPostgreSQL uye kubva kune mamwe dhatabhesi.

Mwedzi wapfuura (Kurume 2017), rutsigiro rwakaonekwa muGrafana. Paunovaka mishumo muGrafana, semuenzaniso, pane mamiriro ehardware yako kana pane mamwe metrics, ikozvino iwe unogona kuvaka iyo yakafanana girafu kana imwe mhando yepaneru kubva kudhata kubva kuClickHouse zvakananga. Izvi zviri nyore, uye tinozvishandisa isu pachedu. Izvi zvinokutendera kuti uwane anomalies. Ndokunge, kana chimwe chinhu chikaitika uye imwe hardware ikadonha kana kuomerwa, saka unogona kutarisa chikonzero kana iyi data yakakwanisa kupinda muClickHouse.

Ndakaona zvichinetsa kunyora mumaturusi aya kana muconsole. Uye ndakafunga kuvandudza yedu yekutanga interface. Uye ini ndakawana zano kubva ChiitikoSQL, SeperSet, Zeppelin.

Waida chii? Ini ndaida kutora mifananidzo, yakagadziridzwa mupepeti, uye kushandisa tsigiro yemaduramazwi ezano. Nekuti ClickHouse ine chinhu chikuru - maduramazwi. Asi zvakaoma kushanda nemaduramazwi, nokuti unoda kuyeuka maitiro ezvinyorwa zvakachengetwa, kureva nhamba kana tambo, nezvimwewo Uye sezvo tichiwanzoshandisa maduramazwi mukusiyana kwavo kwakasiyana, zvakanga zvakaoma kunyora mibvunzo.

Mwedzi mitatu yapfuura kubva pakaburitswa vhezheni yedu yekutanga. Ndakaita anenge mazana matatu nemakumi matatu ekuzvipira kubazi rakazvimiririra uye rakazove Tabix.

Kusiyana neshanduro yapfuura, iyo yainzi ClickHouse-Frontend, ndakafunga kuitumidza zita riri nyore. Uye zvakazoitika Tabix.

Chii chakaonekwa?

Inodhirowa magirafu. Inotsigira ClickHouse SQL syntax. Inopa mazano pamusoro pemabasa uye inogona kuita zvinhu zvakawanda zvinonakidza.

Izvi ndizvo zvinotaridzika seTabix chirongwa. Kuruboshwe kune muti. Pakati pane query editor. Uye pazasi pane mhedzisiro yechikumbiro ichi.

Tevere ini ndinokuratidza kuti mupepeti webvunzo anoshanda sei.

Pano autocomplete inoshanda otomatiki patafura uye inosimudzira, saizvozvo, autocomplete yeminda. Uye mazano pamusoro pemabasa. Kana ukadzvanya ctrl enter, chikumbiro chinoitwa kana kukundikana nekukanganisa. Chikumbiro chakareruka chinotumirwa kuTabix uye mhedzisiro inowanikwa, kureva kuti unogona kukurumidza kushanda neClickHouse.

Maduramazwi, sezvandambotaura, chinhu chinonakidza chatinoshanda nacho zvakanyanya. Uye izvo zvakatibvumira kuita zvinhu zvakawanda. Ngatitii tinochengeta maguta ose mumaduramazwi. Isu tinochengeta chiziviso cheguta uye zita reguta, latitude uye longitude. Uye mune dhatabhesi tinochengeta chete chiziviso cheguta. Saizvozvo, isu tinomanikidza iyo data zvakanyanya.

Ichi chinoita sechinhu chiri nyore, asi chinobatsira muClickHouse nenzira inonakidza. Nekuda kwekuti ClickHouse inongotsigira majoini akaiswa, mubvunzo unokura uchidzika uye wakafara zvakakwana. Uye kana bhuraketi ravhurwa uye kumwe kutaura kwakareba kuchipinda, ipapo chimwe chinhu chiri nyore sekudonhedza mubvunzo chinoita kuti zvive nyore kushanda nemubvunzo pachawo. Nekuti kana muvhunzo wakareba mitsara mazana maviri nemazana matatu uye wakakura kwazvo muhupamhi, zvinobatsira kwazvo kupeta muvhunzo wobva watsvaga imwe nzvimbo kana kuti neimwe nzira.

Muti wechinhu, akawanda uye ma tabo (Vhidhiyo 13:46 https://youtu.be/w1-XsL3nbRg?t=826)

Tevere ini ndinokuratidza nezve muti uye ma tabo. Kuruboshwe kune muti; pamusoro unogona kugadzira akati wandei ma tabo. Matebhu akafanana nenzvimbo yebasa. Iwe unogona kugadzira akati wandei ma tabo uye wotumidza rimwe nerimwe zvakasiyana. Zvakafanana ne mini-system yekuvaka rondedzero.

Matebu anochengetwa otomatiki. Kana iwe ukatangazve bhurawuza rako kana kuvhara kana kuvhura Tabix, zvese izvi zvichachengetwa.

Hotkey - nyore (Vhidhiyo 14:39 https://youtu.be/w1-XsL3nbRg?t=879)

Kune hotkeys uye kune akawanda kwazvo. Ndatora mamwe acho pano semuenzaniso. Uku kushandura ma tabo, kuita chikumbiro kana kuita zvikumbiro zvakati wandei.

Ini ndichakuratidza maitiro ekushanda nemugumisiro. Tinotumira chikumbiro. Pano ndiri kudhirowa chivi, cos uye tg. Unogona kuratidza mhedzisiro, kureva kudhirowa mepu yakajairwa yekoramu. Iwe unogona kujekesa zvakanaka kana zvakaipa maitiro. Kana kungopenda chimwe chinhu chetafura. Izvi zviri nyore kana tafura yakakura uye iwe unofanirwa kuwana imwe inomaly nemeso ako. Pandaitsvaga zvisingaite, ndaisimbisa mimwe mitsara, zvimwe zvinhu zvine girinhi kana tsvuku.

Pane zvinhu zvakawanda zvinonakidza ikoko. Semuenzaniso, maitiro ekukopa muRedmine Markdown. Kana iwe uchida kukopa mhedzisiro pane imwe nzvimbo, izvi zviri nyore. Iwe unogona kungosarudza nzvimbo, taura "Kopira kuRedmine" uye ichakopa muRedmine Markdown kana kugadzira Kupi mubvunzo.

Chinotevera ndeye query optimization. Ndakambokanganwa kutsanangura munda we "date". Uye chikumbiro changu muClickHouse hachina kugadziriswa zvakanyanya, nekukurumidza, asi nekukurumidza, kureva isingasviki sekondi. Pandakaona kuti akadarika mitsetse mingani, ndakatya. Isu hatinyori mitsetse yakawanda kudaro kune iyi tafura muzuva. Ndakatanga kupenengura chikumbiro chacho ndikaona kuti ndainge ndarovha pane imwe nzvimbo. Ndiko kuti, ndakanganwa kuratidza kuti ini handidi data yetafura yese, asi kwenguva yakati.

Tabix ine "Stats" tab, iyo inochengetedza nhoroondo yese yezvikumbiro zvakatumirwa, i.e. ipapo iwe unogona kuona kuti mitsetse mingani yakaverengwa nechikumbiro ichi uye kuti yakatora nguva yakareba sei kuita. Izvi zvinobvumira optimization.

Iwe unogona kugadzira tafura yepivot pamusoro pemubvunzo mhedzisiro. Wakatumira chikumbiro kuna ClickHouse uye wakagamuchira imwe data. Uye ipapo iwe unogona kufambisa iyi data negonzo rako uye kuvaka imwe mhando yetafura yepivot.

Chinhu chinotevera chinonakidza ndechekuronga. Ngatitii tine chikumbiro chinotevera: nokuda kwechivi, cos kubva ku0 kusvika ku 299. Uye kuti uiswedere, unoda kusarudza "Dhira" tab uye iwe uchawana girafu nechivi chako uye cos.

Iwe unogona kupatsanura izvi mumatemo akasiyana, kureva kuti unogona kudhirowa magirafu maviri padivi padivi kamwechete. Nyora murairo mumwe uye wechipiri.

Unogona kudhirowa histograms.

Iwe unogona kuputsa izvi kuita matrix emagirafu.

Unogona kugadzira mepu yekudziya.

Iwe unogona kugadzira karenda yekupisa. Nenzira, ichi chinhu chakanyanya nyore kana iwe uchida kuongorora anomalies pamusoro pegore, i.e., tsvaga kana spikes kana madonhwe. Iyi data visualization yakandibatsira neizvi.

Inotevera ndeye Treemap.

Sankeys chati inonakidza. Iye angave Streamgrahps kana Rwizi. Asi ndinoridaidza kuti Rwizi. Iyo zvakare inobvumidza iwe kuti utarise chero anomalies. Kwakanaka kwazvo. Ndinokurudzira kuishandisa pakutsvaga.

Chinhu chinotevera chinonakidza kudhirowa mepu ine simba. Kana iwe ukachengeta latitude, longitudo mune yako dhatabhesi uye, toti, chengetedza kwaunoenda, kana iwe, semuenzaniso, uine marori kana ndege dzinobhururuka, saka unogona kudhirowa nzira dzekuenda. Zvakare ipapo iwe unogona kuseta kumhanya uye saizi yezvinhu izvi izvo zvavanobhururuka mazviri.

Asi dambudziko remepu iyi nderekuti inodhirowa mepu yenyika chete, hapana ruzivo.

Gare gare ndakawedzera Google mepu. Kana iwe ukachengeta latitude, longitudo, saka unogona kudhirowa mhedzisiro paGoogle mepu, asi pasina tsigiro yendege.

Takakurukura mabasa makuru ekushanda nemhedzisiro uye mibvunzo muTabix.

Iyo inotevera ndeye ongororo yeClickHouse server yako. Kune yakaparadzana "Metrics" tebhu, kwaunogona kuona saizi yedata rakachengetwa rekoramu yega yega. Iyo skrini inoratidza kuti iyi "referrer" ndima inotora inosvika 730 Gb. Kana tikasiya munda uyu, tichachengetedza matatu shards e700 GB imwe neimwe, i.e. nezve 2 TB yatisingade.

Isu tinewo "request_id" ndima yatinochengeta mutambo. Asi kana tikatanga kuichengeta munhamba yenhamba, ndima iyi inodzikira zvakanyanya.

Iyo inoratidzawo sevha yekumisikidzwa uye rondedzero yemanodhi musumbu rako.

Iyo inotevera tebhu ndeye metrics. Ivo vanopinda munguva chaiyo neClickHouse uye vanongobvumidza iwe kuti uongorore mamiriro evhavha uye unzwisise zviri kuitika kwairi. Iyi haisi kutsiva Grafana yakazara. Izvi zvakakosha pakuongorora nekukurumidza.

Iyo inotevera tebhu maitiro. Kubva kwavari unogona kunzwisisa zviri kuitika pane server. Hwisisa zviri kuitika ikoko. Ndakanga ndine chikumbiro chaidya 200 GB yekuverenga nguva yega yega. Ndakaona ichi kutenda kune iyi interface. Ndakamubata ndikamugadzirisa. Uye yakazoita ingangoita 30 GB, i.e. kuita pane dzimwe nguva.

Ndatenda! Uye iri muOpenSource

Ndakapedza. Uye nenzira, ndeye OpenSource, ndeyemahara uye hautombodi kuidhawunirodha. Ivhure mubrowser uye zvese zvichashanda.

Mibvunzo yako

Igor, chii chinotevera? Uchagadzira chombo ichi kupi?

Zvadaro, dashboards ichaonekwa, kureva, zvichida dashboards ichaonekwa. Kubatanidzwa nemamwe databases. Ndakaita izvi, asi handisati ndazviburitsa muOpenSource. Iyi iMySQL uye pamwe PostgreSQL. Ndiko kuti, zvichave zvichiita kutumira zvikumbiro kubva kuTabix kwete kuClickHouse chete, asiwo kune mamwe maturusi.

Zviri pachena kuti basa guru rakaitwa. Yakazova pfungwa yakakwana. Izvi zvakaitwa mubrowser, sezviri pachena, kuitira kubvisa madondoro pamhando dzose dzematemo uye nekukurumidza kukanda chinhu chose pamwechete. Ndanzwa kuti muri on PHP shanda, saka nzira iri nyore ndeyekuinyora mubrowser uye inoshanda kwese kwese. Hapana mibvunzo pamusoro peizvi. Mubvunzo ndouyu. Zvakawanda chaizvo zvakaitwa ikoko. Vangani vanhu vakashanda pane izvi? Uye zvose zvakatora nguva yakareba sei? Nekuti maturusi etsika kazhinji haana basa rakawanda kudaro.

Mumwe munhu kubva kuchikwata chedu akashanda kubva muzhizha kusvika matsutso. Iyi yaive shanduro yekutanga. Ndakabva ndaita 330 commits ndega. Zvamunoona ini nemumwe wangu takazviita nepakati. Mumwedzi mitatu, kubva pashanduro yekutanga kusvika kune yekupedzisira, ndainyanya kuzviita ndega. Asi ini handizivi Javascript zvakanyanya. Iyi yaive yangu chete uye, ndinovimba, yangu yekupedzisira Javascript purojekiti yandakashanda nayo. Ndakaiwana, ndakatarisa-o, kutya. Asi ini ndaida chaizvo kupedzisa chigadzirwa uye izvi ndizvo zvakaitika.

Ndatenda zvikuru nemushumo! Ichi chishandiso chikuru. WITH Tableau Wakaenzanisa here?

Ndatenda. Ndosaka ndakaitumidza kuti Tabix, nekuti mabhii ekutanga akafanana.

Nekuti unokwikwidza?

Pachave neinvestment yakawanda, tichakwikwidza.

Iwe unogona sei kupa kutengesa kune vemukati vanoongorora kuti chishandiso ichi chichatsiva zvachose *Tafura*? Zvii zvichava nharo?

Inoshanda natively neClickHouse. Ndakaedza Tableau, asi haugone kunyora tsigiro yemaduramazwi uye nezvimwe zvakadaro. Ndinoziva mashandiro anoita vanhu neTabix. Ivo vanonyora mubvunzo, voisa kuCSV uye voisa kuBI. Uye vari kutoita chimwe chinhu ikoko. Asi ini ndine nguva yakaoma yekufungidzira kuti vanoita sei izvi, nekuti i graphical chishandiso. Inogona kuburitsa mitsara ye5, yakakura ye000 mitsara, asi kwete zvakare, zvikasadaro browser haizogone.

Ndiko kuti, pane zvimwe zvakakomba zvisingakwanisi pahuwandu hwe data, handiti?

Ehe. Handigone kufungidzira kuti iwe ungada kurodha zviuru gumi mitsara mutafura yako pane yako browser skrini. Yechii?

Izvi zvinoreva here kuti iyi ndiyo interface yekukurumidza kuona data? Kumonyorora zvishoma, kumonyorora?

Hongu, kurumidza kuona kuti inoshanda sei uye ingovaka pfupiso girafu. Uye wozvipa kumwe. Isu tine yedu yekushuma system, kubva kwandinongotora ichi chikumbiro. Ini ndinodhirowa muTabix ndotumira kune yedu yekushuma.

Uye mumwe mubvunzo. Cohort analysis?

Kana pane zvikumbiro, tichazviwedzera.

Wakatanga rinhi kuishandisa? ClickHouse, kushandiswa kwakatora nguva yakareba sei? DzvanyaImba uye kuunza ku production state?

Sezvandakataura, takaita bvunzo sumbu munguva pfupi pfupi. Takaisa mukati memazuva maviri. Uye takaiedza kwemamwe mavhiki akati wandei. Uye takasvika pakugadzirwa mumwedzi 3. Asi isu takanga tine ETL yedu, i.e. chishandiso chekurekodha data. Uye akanyora mune zvose zvaaigona. Anokwanisa kunyora muMongoDB, Cassandra, MySQL. Zvaive nyore kumudzidzisa kunyora muClickHouse. Takanga tine zvivakwa zvakagadzirirwa zvekugadzirisa nekukurumidza. Mukati memwedzi mitatu takatanga kukanda chikamu chekutanga. Mumwedzi mitanhatu takasiya zvachose zvimwe zvese. Isu tine imwe chete ClickHouse yasara.

Igor, ndinokutendai zvikuru nemushumo. Ndaifarira chaizvo kushanda kwekuvaka nzira uchishandisa mepu. Pane here zvirongwa zvekubatanidza neYandex.Maps uye kunyanya neYandex.Maps?

Ndakaedza kubatanidza panzvimbo yeGoogle mepu, asi handina kuwana rima dingindira paYandex.Maps. Handina kukuudza chidimbu chimwe chete. Ndichadzosera kumashure kuwedzera.

Slide – Google mepu. Pane murairo "DRAW_GMAPS", unodhirowa mepu. Pane murairo "DRAW_YMAPS", i.e. inogona kudhirowa Yandex.Map. Asi chokwadi, pasi pemurairo uyu pane Javascript, i.e. iyo data yaunogamuchira kubva kuClickHouse inogona kuendeswa kuJavascript, yaunonyora pano. Uye iwe une nzvimbo yekubuda kwainofanira kudhonzwa. Unogona kudhirowa chero girafu, kureva chero girafu, mepu, unogona kudhirowa chako chikamu. Izvi zvisati zvaitika, ndaive neimwe raibhurari yekudhirowa iwo magirafu pachawo.

Ndiko kuti, pane chishandiso chekugadzirisa mashandiro ekuratidzira?

Chero. Iwe unogona kutora uye kudzorerazve mavara aya, uchiita kuti asatsvuka, asi bhuruu, girini.

Ndatenda neshumo! Wanga uine siraidhi yairatidza mamwe maturusi emibvunzo DzvanyaImba yekuvaka dashboards uye analytical report. Ndinonzwisisa kuti panguva yawakatanga kushanda nayo ClickHouse, hapana maadapter akanyorerwa maturusi aya. Uye ndiri kushamisika kuti sei wakafunga kugadzira chako chishandiso, pane kunyora adapta yeimwe yakagadzirira-yakagadzirwa chishandiso? Ini ndinofunga kuti kugadzirisa iyo test editor inokurumidza. Nei wakasarudza kuita basa rakawanda kudaro?

Pane chinhu chinonakidza pano - chokwadi ndechekuti ini ndiri technical director, kwete data sainzi. Panguva yatakatanga kushandisa Druid, mepu yangu yemugwagwa yaive ne50% yemabasa - ngativerengei izvi, kana kuti ngativerengei izvi, kana kuongorora izvi. Uye zvakazoitika kuti isu takaita ClickHouse. Uye akatanga kukurumidza kuvaka zvese, kuverenga, uye nekukurumidza akavhara nzira yake. Uye panguva iyoyo ndakaona kuti ndakanga ndisina ruzivo muData Science uye data visualization. Tabix imhando yebasa rangu rechikoro rekudzidza kuona data. Ini ndanga ndichitarisa nzira yekuzadzisa Zeppelin. Ini handifarire hurongwa hwake. Redash ndakatarisa kuti ndoiwedzera sei, but a normal editor yaindikwanira. Uye SuperSet yakanyorwawo mumutauro wandisingade chaizvo. Uye saka ndakafunga kutenderera, uye izvi ndizvo zvakaitika.

Igor, unogamuchira zvikumbiro zvePull?

Hongu.

Ndatenda zvikuru nemushumo! Uye mibvunzo miviri. Kutanga, iwe hausi kutaura zvinobata kumeso nezvazvo Javascript. Wakanyora muJavascript isina chinhu kana kuti imhando yemhando?*

Zvirinani muJavascript isina chinhu.

Saka chii framework?

Angular.

Zviri pachena. Uye mubvunzo wechipiri. Wambofunga here R и *Shiny**?*

Kufunga nezvazvo. Yakatamba.

Unogonawo kunyora adapta.

Ari. Zvinoita sekuti nharaunda yakazviita, asi, sezvandakapindura mubvunzo wapfuura, ndaida kuzviedza pachangu.

* Kwete, maererano nekuona, zviripo zvakare.

Iwe unoti kune chinhu chakadaro uye chichakukwevera iwe girafu. Ndakavhura bhuku rekuona data. Uye ndakafunga: “Rega ndiedze kuona iyi data. Ndichamunyorera kuti agadzirise data." Uye ini ndakatanga kunzwisisa zviri nani data feed tekinoroji. Uye dai ndakanga ndatora chikamu chakagadzirirwa-chakagadzirwa, ini pachangu ndingadai ndakadzidza zvakanyanya kuishandisa, ndiko kuti, kuona. Asi hongu, ndaifarira R, asi handisati ndaverenga bhuku "R for Dummies" parizvino.

Ndinokutendai!

Mubvunzo wakapfava. Pane dzimwe nzira dzekukurumidza kurodha chikwangwani kana kuronga?

Inogona kuiswa kuCSV kana Excel.

Kwete data, asi ndiro yakagadzirira, girafu rakagadzirirwa? Somuenzaniso, kuratidza mukuru.

Pane bhatani "Rodha" uye pane bhatani "Isa girafu mu png, mu jpg".

Ndinokutendai!

PS Mini-mirayiridzo yekuisa tabix

• Download ichangoburwa
• Unpack, kopi dhairekitori build mu nginx midzi_path
• Gadzirisa nginx

Source: www.habr.com