Question

My query looks like:

SELECT * 
  FROM dbo.TestTable TT
  JOIN dbo.fnListParseAndSplit('test,tt,zz,er,ts',',') L ON TT.Name like '%' + L.ListMember + '%'

My issue is that on the execution plan I always get Index Seek for both cases when joining on:

1. T.Name like ‘%’ + L.ListMember + ‘%’

2. T.Name like L.ListMember + ‘%’.

I was of the opinion that they are completely different and when the join is like ‘%’ + L.ListMember + ‘%’ I should get an index scan.
Do I get something wrong on the execution plan? Or the SQL server 2008 is so smart that it can optimize my query?

I attached the execution plan. Do you know why do I have an index seek on the TestTabe Name column? The name column have an unique index … but in this case I would expect a scan instead of seek.

http://softrun.ro/executionplan.png

Answer 1

Index seek is related first part of join, if you look at the posted plan, you’ll see table scan for output of your parse function.