Question

I’ve been wondering, what’s the complexity difference between these two cases:

Case 1: An OLE DB Source query, written like this:

 select *
 from A
 where A.value > 1

Case 2: An OLE DB Source query, and the where clause is put in Conditional Split

 select * from A

with the Conditional Split after the OLE DB Source query, containing: value > 1.

Performance-wise, does it have any difference? And for more complex query, does it have any significant impact also?

Answer 1

Yes there is a performance difference.

Case 2 will return all data records from table A and store them in the SSIS memory buffer before passing them on to the Conditional Split Component for filtering.

Case 1 immediately returns a smaller data set to the SSIS memory buffer.

For further reading take a look at:

• Understanding SSIS Data Flow Buffers video
• Improving Performance of the Data Flow
• Monitoring Performance of the Data Flow Engine