Question

How should i go about doing something like:

IF the following select has any rows

SELECT ID, NAME FROM TABLE WHERE ID=@ID

then return that actual select, otherwise use another select statement.

I use somwthing like

IF EXISTS(SELECT ID, NAME FROM TABLE WHERE ID=@ID)
    SELECT ID, NAME FROM TABLE WHERE ID=@ID
ELSE
    SELECT QTY FROM TABLE WHERE ID=@ID

but this way i use the first select twice(once in the check and once the actual return record set.

Answer 1

It’s not as inefficient as you think. The EXISTS keyword means that the optimiser stops after 1 record is found.

But, if you really want to avoid repetition, you can do something like…

SELECT ID, NAME FROM TABLE WHERE ID=@ID

IF (@@rowcount = 0)
    SELECT QTY FROM TABLE WHERE ID=@ID

The problem here is that you now return up to 2 record sets to the client. The first of which my be empty (but field headers, etc, are still returned – just with 0 records).

You could avoid that by inserting results into a table variable, and only returning them to the client if there are any records. But I think it would be noticeably slower than just using EXISTS() like you already do.