Subject | RE: [ib-support] SELECT in stored proc returns wrong value |
---|---|
Author | Ann W. Harrison |
Post date | 2001-06-18T18:39:11Z |
At 11:43 PM 6/17/2001 +0200, Christian Gütter wrote:
nothing doesn't change the variables used in the INTO clause. If
they were null to start, they'll still be null. If not, not.
Nor is the behavior of SUM a bug. The sum of zero rows is null.
Intuitive or not, that's the behavior required by the SQL standard.
As for doing a "select count (*) ... " before doing a select or
update, that's generally a terrible idea because it doubles the
amount of work done. If you expect the count to be in the range
0-10, not so bad, but still, it's much better to count the results
in the stored procedure.
Regards,
Ann
www.ibphoenix.com
We have answers.
>Interbase behaves differently in two situations which are almost theSorry, no, it's not a bug. As Helen said, a select that returns
>same.
>
>1) When doing a SELECT outside a stored proc, it returns NULLs.
>
>2) When doing the same SELECT inside a stored proc, it leaves the result
> variables untouched, they are not NULL.
>
>For me, this is not consistent behaviour.
>So one might call it a bug, not a serious one, but still a bug.
nothing doesn't change the variables used in the INTO clause. If
they were null to start, they'll still be null. If not, not.
Nor is the behavior of SUM a bug. The sum of zero rows is null.
Intuitive or not, that's the behavior required by the SQL standard.
As for doing a "select count (*) ... " before doing a select or
update, that's generally a terrible idea because it doubles the
amount of work done. If you expect the count to be in the range
0-10, not so bad, but still, it's much better to count the results
in the stored procedure.
Regards,
Ann
www.ibphoenix.com
We have answers.