你問問題不清楚

你問問題不清楚

第一個笨辦法是: perform a parallel full table scan in the largest table with a hint like /* full(x) parallel(x, 16)*/. It depends on your server's capacity,you can set different parallel numbers. Then you can joint this table with others by index.

如果你有小一點的table,you can load it into memory by using hash joint. Just check Oracle SQL tuning help.

偶爾路過,幫不了忙。隨嘴帶兩句。BTW, my rate is $250US/Hr.

If you can not handle these tables with records less than millions, better you....

請您先登陸,再發跟帖!