On Tuesday, February 9, 2016, Vitaly Burovoy <
vitaly.burovoy@gmail.com> wrote:
On 2/9/16, Harald Fuchs <hari.fuchs@gmail.com> wrote:
> Johannes <jotpe@posteo.de> writes:
>
>>> What the reason to execute all statements which return different
>>> columns at once?
>>>
>>>> Saving roundtrips,
>>>
>>> In most cases they are not so big. Getting a bunch of duplicated data
>>> is wasting you network bandwidth and don't increase speed.
>>
>> In my and your example no duplicated data (result sets) is send over the
>> network. The server do not need to wait until the client snips out the
>> id and sends it id in the next query again. So the server can compute
>> the result set without external dependencies as fast as possible.
>
> Sounds like what you're really after is a stored procedure, isn't it?
Unfortunately, his case is different, because he needs to get two
different set of rows that is impossible even with stored procedures.
Correct, though it might be workable to use cursors in this situation. Not exactly sure how, though...
David J.