尚硅谷大數(shù)據(jù)Hive 3.x教程全新升級版(基于hive3.1.3)

Hive 優(yōu)化階段 map join 優(yōu)化,我用我們公司的測試集群跑了一下,執(zhí)行的時間如下,供大家參考
------------------ MAP JOIN 測試 set hive.execution.engine=mr; --- 未開啟 MAP JOIN --- execution: 3 m 48 s 547 ms, fetching: 234 ms set hive.auto.convert.join=false; select * from order_detail od join product_info product on od.product_id = product.id join province_info province on od.province_id = province.id; --- 方案一:開啟 map join set hive.auto.convert.join=true; set hive.auto.convert.join.noconditionaltask=false; set hive.mapjoin.smalltable.filesize=25285707; --- execution: 3 m 59 s 70 ms, fetching: 201 ms select * from order_detail od join product_info product on od.product_id = product.id join province_info province on od.province_id = province.id; -- 方案二:開啟 map join set hive.auto.convert.join=true; set hive.auto.convert.join.noconditionaltask=true; set hive.auto.convert.join.noconditionaltask.size=25286076; -- execution: 1 m 28 s 499 ms, fetching: 151 ms select * from order_detail od join product_info product on od.product_id = product.id join province_info province on od.province_id = province.id; -- 方案三:開啟 map join set hive.auto.convert.join=true; set hive.auto.convert.join.noconditionaltask=true; set hive.auto.convert.join.noconditionaltask.size=25285707; -- execution: 3 m 26 s 115 ms, fetching: 165 ms select * from order_detail od join product_info product on od.product_id = product.id join province_info province on od.province_id = province.id;
標簽: