Hello
I have a table with 200 million rows (tableA) and an unique non-clustered index (ix_A) on id1 and id2.
when I run a simple query
select b.id1,b. id2, b.otherfield1, b.otherfield2, b.otherfield3 from #temp a join tableA b on a.id1 = b.id1
The temp table has 15000 rows and the execution plan having a index seek on ix_A. However, when I run same query, but temp table rows increased to 30000, the execution plan change to a table scan on tableA , of cause the running time has increased significantly. Can anyone tell me why? and any advice how to tune it, the table scan execution plan didn't show any missing index on top of execution plan.
Any help would be much appreciated,
Carqu