I have been able to setup statistics so that federator engine identifies the smaller table T1 & the larger Table T2 among the 2 tables. And when clicked "Explain Plan" it shows that it would use "Semi-Join" to fetch results of T2 using the joined column from T1. And it's using I, P as strategies. It looks perfect with this plan.
But when it's starts executing - it rightly starts with executing the smaller table T1 & the larger table T2 starts once the T1 is completed.
the only problem is - it's still doing a full scan of the larger table T2 ( i was expecting that it would use the result set fetched from T1)
have anyone of seen this behavior - what could be missing here ? please suggest. Thanks