数据库大表小表

The "select * from big where object_id in ( select object_id from small )" will sort BIG once and SMALL once and join them (sort merge join) in all likelyhood.

But: The "select * from big where exists ( select null from small where  small.object_id = big.object_id )" will tend to FULL SCAN big once and FOR EVERY   ROW in big -- FULL SCAN small