Question

I just set an SQL Stored Procedure to return a specific row in a specific order (from more to less important):

-- 1st Level -> Query a more detailed object...
SELECT ... WHERE something
-- if a result is find, the SP is returning the correct row.
-- if rowcount = 0 then, the SP is returning an empty row 
-- and continues with the next query (less specific than the 1st one).

IF @@ROWCOUNT = 0
BEGIN
SELECT .... WHERE something else... 
 -- Again, if a result exists, the SP returns the correct row, but also 
 -- the result of the 1st query is returned without rows...
 IF @@ROWCOUNT = 0
 BEGIN
 SELECT .... WHERE something else again...
 END
END

Is there any option to only return the SELECT statement that returns a non empty row? I don’t want to return empty rows… and each time the row is an “sub-level” of the 1st query, I get empty results before the correct row.

I was thinking to create a table variable. Are there any better ways?

Answer 1

Since you are looking for the lack of rows why not use an if/else statement in your procedure. Something like:

If exists(/*firstQuery*/ SELECT ... WHERE something) begin
   SELECT ... WHERE something
end else if exists (/*secondQuery*/ SELECT .... WHERE something else... ) begin
   SELECT .... WHERE something else... 
end else /*Third query*/ begin
    SELECT .... WHERE something else again...
end

Just remember when checking is something exists you should select only one column (for example the id) as if you have calculated columns then this can slow down your procedure.