倉庫的主要技術(shù)、以AnalyticDB為代表的云原生數(shù)據(jù)倉庫的核心技術(shù)、 6 6 9 云原生數(shù)據(jù)倉庫AnalyticDBMySQL1.產(chǎn)品簡介云原生數(shù)據(jù)倉庫AnalyticDBMySQ2.產(chǎn)品定位云原生數(shù)據(jù)倉庫AnalyticDBMySQ3.產(chǎn)品特點與優(yōu)勢4.產(chǎn)品核心功能云原生數(shù)據(jù)倉庫AnalyticDBMySQ1.產(chǎn)品架構(gòu)云原生數(shù)據(jù)倉庫AnalyticDBMyS用。2.優(yōu)化器介紹云原生數(shù)據(jù)倉庫AnalyticDBMyS3.彈性計算層云原生數(shù)據(jù)倉庫AnalyticDBMySnationkey=lineitem.例。云原生數(shù)據(jù)倉庫AnalyticDBMyS4.存儲層云原生數(shù)據(jù)倉庫AnalyticDBMyS云原生數(shù)據(jù)倉庫AnalyticDBMyS行列混存是一種以列存為基礎(chǔ)兼顧行存的模式，類似于Hadoop中的云原生數(shù)據(jù)倉庫AnalyticDBMyS1.數(shù)據(jù)庫對象及操作云原生數(shù)據(jù)倉庫AnalyticDBMyS一份。云原生數(shù)據(jù)倉庫AnalyticDBMyS云原生數(shù)據(jù)倉庫AnalyticDBMySQL版解析與實踐（上）192.權(quán)限與數(shù)據(jù)安全云原生數(shù)據(jù)倉庫AnalyticDBM務(wù)。云原生數(shù)據(jù)倉庫AnalyticDBM3.智能索引云原生數(shù)據(jù)倉庫AnalyticDBM云原生數(shù)據(jù)倉庫AnalyticDBM4.數(shù)據(jù)存儲冷熱分離云原生數(shù)據(jù)倉庫AnalyticDBM云原生數(shù)據(jù)倉庫AnalyticDBM參考table_usage表字段信息參考table_usage表字段信息云原生數(shù)據(jù)倉庫AnalyticDBMySQL版解析與實踐（上）265.物化視圖?物化視圖是數(shù)倉領(lǐng)域的核心特性之一。不同于邏輯視圖（view物化視圖CREATECREATEMATERIALIZEDVIECLUSTEREDKEYINDEX(n)云原生數(shù)據(jù)倉庫AnalyticDBM6.備份恢復(fù)云原生數(shù)據(jù)倉庫AnalyticDBM云原生數(shù)據(jù)倉庫AnalyticDBM1.彈性能力：資源池?用戶自定義創(chuàng)建的資源池，可以進(jìn)行資源池的計算資源、查詢執(zhí)行模式、CREATECREATERESOURCEGROUP2.彈性能力：EIU瓶頸。3.彈性能力：分時彈性4.操作演示-實現(xiàn)AnalyticDBMySQL版 1.數(shù)據(jù)導(dǎo)入導(dǎo)出3.DML外表式數(shù)據(jù)導(dǎo)入導(dǎo)出CREATECREATETABLEIFNOTEXIS)"url":"jdbc:mysql://mysql-vpENGINEENGINE='ODPS'---存儲引擎是MaxCompute4.數(shù)據(jù)同步倉遷移。建表。 1.查詢流程和執(zhí)行計劃2.算子操作。3.影響查詢性能的因素4.典型慢查詢5.慢查詢診斷與優(yōu)化6.診斷與優(yōu)化示例/*+LEFT_TO_RIGHT_ENABLED=true/*+LEFT_TO_RIGHT_ENABLED=true，CASCADES_OPTIM/*+filter_not_pushdown_columns=[test01.7.優(yōu)化原則新操作。云原生數(shù)據(jù)倉庫AnalyticDBMy1.批量打包方式寫入數(shù)據(jù)2.更新數(shù)據(jù)3.刪除數(shù)據(jù)/分區(qū)/表云原生數(shù)據(jù)倉庫AnalyticDBMy?數(shù)據(jù)刪除頻率低、基于主鍵為條件的刪除，可通過DELETEFROMWHE4.批量導(dǎo)入數(shù)據(jù)5.實時導(dǎo)入數(shù)據(jù)云原生數(shù)據(jù)倉庫AnalyticDBMy云原生數(shù)據(jù)倉庫AnalyticDBMy1.注冊阿里云賬號2.阿里云賬號實名認(rèn)證1.前提條件2.創(chuàng)建數(shù)倉版（3.0）集群1.數(shù)據(jù)庫賬號類型據(jù)庫。2.創(chuàng)建高權(quán)限賬號1.背景信息2.數(shù)倉版（3.0）設(shè)置白名單 1.背景信息2.使用DMS連接云原生數(shù)據(jù)倉庫AnalyticDBMySQL版1.背景信息 1.前提條件通過以下步驟在對象存儲（ObjectStorageService，簡稱OSS）中創(chuàng)建存儲0006;hello_world_62.操作步驟?對于數(shù)據(jù)量較大的CSV或TEXT數(shù)據(jù)文件，強(qiáng)烈()功能演示二：創(chuàng)建數(shù)據(jù)倉庫并完成數(shù)據(jù)同步63classbigint(20)NOTNULL,namevarchar(32)NOTNULL,update_NOTNULLDEFAULTCURREN功能演示二：創(chuàng)建數(shù)據(jù)倉庫并完成數(shù)據(jù)同步644.創(chuàng)建外部映射表，使用以下命令在目標(biāo)庫adb_demo中創(chuàng)建一張名為CREATECREATETABLEIFNOTEXISTS)"url":"jdbc:mysql://mysql-vpc5.創(chuàng)建目標(biāo)表，使用以下命令在目標(biāo)數(shù)據(jù)庫adb_demo中創(chuàng)建一張名為))功能演示二：創(chuàng)建數(shù)據(jù)倉庫并完成數(shù)據(jù)同步65REPLACEREPLACEINTOmysql_imp云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQL1.產(chǎn)品簡介圖。云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQL2.發(fā)展歷程3.產(chǎn)品定位云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQL4.產(chǎn)品特點和優(yōu)勢5.核心功能AsdefinitionAs'objfile，Linksymbol'...云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQL云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQL云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQL6.產(chǎn)品選型1.產(chǎn)品架構(gòu)云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQL2.模塊組件發(fā)協(xié)調(diào)。云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQL云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQL3.數(shù)據(jù)模型云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQL1.數(shù)據(jù)庫對象及操作云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQL云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQL版解析與實踐（上）77象。）：云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQL版解析與實踐（上）78CREATECREATETABLEreplic云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQL云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQL版解析與實踐（上）80DEFAULTSUBPARTITION云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQL版解析與實踐（上）81WITH(APPENDONLY=TRUE，ORIENTATION=COLUMN)DISTRIBUTEDBY(a云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQL?調(diào)優(yōu)過程中，從執(zhí)行計劃中看到表行數(shù)估算為1行，計劃中出現(xiàn)較多的generate_series(1，10O_ORDERSTATUSCHAR(1)O_TOTALPRICEDECIMAL(15，2)O_ODRDEAPRIDRITYCHAR(15)O_SHIPPRIORITYINTEGERO_COMMENTVARCHAR(79))O_ORDERSTATUSCHAR(1)O_TOTALPRICEDECIMAL(15，2)O_ODRDEAPRIDRITYCHAR(15)云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQLO_SHIPPRIORITYO_SHIPPRIORITYINTEGERO_COMMENTVARCHAR(79))WITH(APPENOONLY=TRUE,2.權(quán)限管理云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQL3.UDF與存儲過程云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQL云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQLCREATECREATEORREPLACEFU 云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQLUPDATEmytabsETfirstname:three'WHERElastnEXCEPTIONWHENOTHERS云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQLcreatecreateoperator>!<(procedure=f_s云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQL4.備份恢復(fù)云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQL云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQL5.資源隔離云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQL制。ACTIVE_STATEMENTS=integ如果沒有顯式地把角色指派資源隊列，它們將進(jìn)入默認(rèn)資源隊列=#=#ALTERROLEnameRESOU=#CREATEROLEnameWITHLOG云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQL=#=#AITERRESOURCEQUEUEadhocWITH(ACIIVE_STATESELECT*FROMSELECT*FROMgp_toolkit.g如果想要持續(xù)跟蹤資源隊列的統(tǒng)計信息和性能，用戶可以使用云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQLpgroles.rolresqueue=gptoolkit.gpresqueuestatus.queSELECTSELECT*FROMgp_toolkit.gp_resq_p云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQLselectselectdatname,pg_sizepg_catalog.pg_databselectselectpg_size_pretty(pg_databselectselecttable_namefrominform云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQLselectselecttable_name,pg_total_reinformation_schema.tabselectselectpg_size_pretty(pg_relation_siselectselectpg_size_pretty(pg_total_relation_sifrompg_namespacewhere云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQL云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQL2.數(shù)據(jù)同步監(jiān)控-用戶側(cè)云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQL版 3.鏈路典型問題排查1.控制臺監(jiān)控云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQL版2.云監(jiān)控預(yù)警云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQL版1.Cascade框架SQL優(yōu)化器2.向量化計算引擎云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQL版4.執(zhí)行計劃兩種收集模式信息。云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQL版5.執(zhí)行計劃6.可視化執(zhí)行計劃云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQL版PostgresPostgresEXPLAINVisualizer7.如何發(fā)現(xiàn)問題8.通過索引提升查詢性能云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQL版9.消除RedistributeMotion云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQL版10.避免下盤11.鎖的檢測及處理云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQL版selectselect*fromgp_toolkit.gp_locks_on_rel12.空間回收云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQL版13.避免數(shù)據(jù)傾斜云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQL版selectselectgp_segment_id,pg_size_pretty(pg_total_relationsizeltable小表（總行數(shù)低于1萬）優(yōu)先選擇復(fù)制表分布策略（DISTRIBUTED若沒有數(shù)據(jù)分布均勻的字段作為分布鍵使用，采用隨機(jī)分布策略云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQL版（）較大的是()?云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQL版（） 云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQL1.操作步驟云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQL云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQL1.設(shè)置白名單操作步驟2.AnalyticDBPostgreSQL連接方式3.使用ECS部署客戶端方式連接AnalyticDBPostgreSQL云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQLtartar-xzvfadbpg_client_package.el7.x86_6psqlpsql"host=yourgpdbaddress.gpdb.ruser=gpdbaccountpa云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQL1.建表并插入數(shù)據(jù)CREATECREATEORREPLACEFUNCTIONgenerate_series(1,云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQL 1.檢查實例的可用性2.查看當(dāng)前運(yùn)行的SQL3.終止運(yùn)行中SQL4.查看節(jié)點狀態(tài)信息云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQL5.查看節(jié)點狀態(tài)是否有down6.查看數(shù)據(jù)庫大小selectselectpg_size_pretty(pg_datab7.查看表執(zhí)行過的歷史操作8.查看哪些表被鎖1.前置條件2.操作步驟該鏈路。云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQL?目標(biāo)端信息，選擇已創(chuàng)建好的ADBPG實例，選擇實例類型“Analytic云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQL云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQL1.前置條件2.操作步驟云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQL云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQL 1.前置條件2.操作步驟云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQL云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQL云原生數(shù)據(jù)倉庫AnalyticDBPostgreSQL 云上數(shù)據(jù)倉庫的架構(gòu)方案1291.用戶行為日志基本流程簡介云上數(shù)據(jù)倉庫的架構(gòu)方案1302.用戶行為日志分析解決方案與收益云上數(shù)據(jù)倉庫的架構(gòu)方案131層。云上數(shù)據(jù)倉庫的架構(gòu)方案132?基于DMS實現(xiàn)一站式全鏈路數(shù)據(jù)管理與服務(wù)。1.案例背景云上數(shù)據(jù)倉庫的架構(gòu)方案1332.AnalyticDB的BI實時數(shù)據(jù)分析加速方案3.AnalyticDB采用數(shù)據(jù)同步獲取最新業(yè)務(wù)數(shù)據(jù)云上數(shù)據(jù)倉庫的架構(gòu)方案1344.AnalyticDB的實時數(shù)據(jù)分析引擎加快實時查詢5.AnayticDB如何支撐BI實時數(shù)據(jù)云上數(shù)據(jù)倉庫的架構(gòu)方案135需求。1.用