Roy,

> One thing I would definitely suggest is that you name your working
variables/parameters differently to your fields. Tim's parser may be
able to figure them out but it sure confuses me. It would also make the
code more readable, at least in my opinion.
>

You are absolutely right, Roy.

Regards Uli

Uli,

<< The execution time of this procedure is about 4 sec! though only one row
is updated (Indexes on Mahnstufe and RechnungenID in Mahnungen are set
properly). When I replace the paramaters by a hardcoded ID, it runs twice as
fast, which I don't understand at all. >>

There should be no difference between a hard-coded value and a simple
parameter value in terms of performance.  Are you sure that the parameter
values being passed in are a) not NULL and b) the values that you think they
are ?

--
Tim Young
Elevate Software
www.elevatesoft.com

Tim,

<<
There should be no difference between a hard-coded value and a simple
parameter value in terms of performance.  Are you sure that the parameter
values being passed in are a) not NULL and b) the values that you think they
are ?
>>

I guessed that there should be no difference. I tried the whole thing again very thoroughly.

Result: SP using a hardcoded (existing and correct value): 0.8 sec.
          SP using a param (integer: existing and correct value): 2.3 sec.

As the procedures are quite long, I attached both in a textfile. I tested all in EDBManager.

Regards Uli

Uli,

<< I guessed that there should be no difference. I tried the whole thing
again very thoroughly.

Result: SP using a hardcoded (existing and correct value): 0.8 sec.
          SP using a param (integer: existing and correct value): 2.3 sec.

As the procedures are quite long, I attached both in a textfile. I tested
all in EDBManager. >>

There's not much I can see here without the data to execute it against.  Is
the database particularly large ?  If not, could you email it to me so I can
test with the same data ?

Thanks,

--
Tim Young
Elevate Software
www.elevatesoft.com

Tim,

> There's not much I can see here without the data to execute it against.
> Is the database particularly large ?  If not, could you email it to me so
> I can test with the same data ?

OK, I see if I can zip it.

Uli

Uli,

<< OK, I see if I can zip it. >>

Thanks, I've found the problem and a fix will be in 1.09, which looks like
it will finally make it out the door today.

--
Tim Young
Elevate Software
www.elevatesoft.com