Viewing 15 posts - 511 through 525 (of 582 total)
quote:
But I wonder how many people have thought about making a single Phone Numbers table with maybe a fieldtbl_From
to know which table...
Tim Wilkinson
"If it doesn't work in practice, you're using the wrong theory"
- Immanuel Kant
September 17, 2003 at 3:09 am
code please?
Tim Wilkinson
"If it doesn't work in practice, you're using the wrong theory"
- Immanuel Kant
September 16, 2003 at 4:44 am
I would favour using a lookup table even for these small domains. Although you are right that check constraints would would involve fewer joins (including the SQL Server internal joinlike...
Tim Wilkinson
"If it doesn't work in practice, you're using the wrong theory"
- Immanuel Kant
September 16, 2003 at 2:55 am
Dave, just had a thought - if it's practical to do on a test environment, you could use DBCC pintable, which forces the whole table to be RAM resident, so...
Tim Wilkinson
"If it doesn't work in practice, you're using the wrong theory"
- Immanuel Kant
August 6, 2003 at 7:42 pm
So you've eliminated the (apparently strong) possibility that the initial delay is due to building the query plan??????
Tim Wilkinson
"If it doesn't work in practice, you're using the wrong theory"
- Immanuel Kant
August 6, 2003 at 2:33 am
I wrote 'procedure cache' - meant 'plan cache'.
Tim Wilkinson
"If it doesn't work in practice, you're using the wrong theory"
- Immanuel Kant
August 3, 2003 at 5:33 pm
It sounds like it is the query plan. If your app is sending simple textual SQL statements with constants as SARGs, then there won't be a match in the procedure...
Tim Wilkinson
"If it doesn't work in practice, you're using the wrong theory"
- Immanuel Kant
August 3, 2003 at 5:29 pm
ADO spawns additional connections! May not be the cause of your problem, but is interesting anyway...
http://support.microsoft.com/default.aspx?scid=kb;EN-US;194979
Tim Wilkinson
"If it doesn't work in practice, you're using the wrong theory"
- Immanuel Kant
July 23, 2003 at 3:34 am
I (and Guarddata, it appears) still don't believe that the code as you describe it would open a new connection. Something else is going on in the code.
You might...
Tim Wilkinson
"If it doesn't work in practice, you're using the wrong theory"
- Immanuel Kant
July 2, 2003 at 10:09 am
Using nolock means that transactions are no longer honoured. You can read data which has not been (and might never be) committed. Equally you might be unable to see data...
Tim Wilkinson
"If it doesn't work in practice, you're using the wrong theory"
- Immanuel Kant
July 2, 2003 at 4:22 am
Yes - agree with GuardData - closing the rocordset ought to release any locks it is holding - in some configurations of the lock- and cursor- type parameters, the rst...
Tim Wilkinson
"If it doesn't work in practice, you're using the wrong theory"
- Immanuel Kant
July 1, 2003 at 4:58 pm
I suspect that FK caused the problem because the new connection is waiting for the other connection to commit and release its locks so that the new conn. can check...
Tim Wilkinson
"If it doesn't work in practice, you're using the wrong theory"
- Immanuel Kant
July 1, 2003 at 4:46 pm
Don't understand what's happening here, but here's a few randomly selected musings:
Have you tried updating the table from Query analyser rather than the app?
How about dropping the trigger and then...
Tim Wilkinson
"If it doesn't work in practice, you're using the wrong theory"
- Immanuel Kant
July 1, 2003 at 12:27 pm
Is there any reason why you wouldn't remove the mistake of creating a new connection? That would solve your problem.
A couple of subsidiary points:
Using @@identity is more efficient than select...
Tim Wilkinson
"If it doesn't work in practice, you're using the wrong theory"
- Immanuel Kant
July 1, 2003 at 11:33 am
alzdba - how do you see the trans. iso. level via Profiler? That would be very convenient...
Tim Wilkinson
"If it doesn't work in practice, you're using the wrong theory"
- Immanuel Kant
July 1, 2003 at 8:15 am
Viewing 15 posts - 511 through 525 (of 582 total)