Viewing 15 posts - 106 through 120 (of 249 total)
An order by expression can be a non-negative integer to represent an expression in the select list; however, I expect that an expression that evaluates to an integer is just an integer, not...
August 22, 2006 at 10:39 am
Is this related to notification services? There is also "set statistics time on" and "set statistics io on" which can be handy.
August 22, 2006 at 10:11 am
Are you really sure that space is not a problem? The maintenance plan can take a lot of space when doing disk backups and reindexing - it's not smart about...
August 21, 2006 at 6:24 pm
Did the index maintenance get overlooked? Have you checked for index fragmentation?
August 21, 2006 at 4:53 pm
I've created a job that logs the stats every 10 minutes and a report that lists the delta for each 10 minute duration. Here is sample output...
August 21, 2006 at 12:57 pm
In my case, I see CXPACKET waits while rebuilding indexes. The other CXPACKET waits could well be due to design and code. However, a setting change or hardware update is a lot...
August 21, 2006 at 11:18 am
Why wouldn't David's version work? Nulls? Duplicates?
SELECT Country_Name
FROM tblCountry
WHERE NOT Country_Id IS NULL
ORDER BY
CASE Country_Id WHEN 26 THEN 1 ELSE 2 END,
Country_Name
SELECT...
August 18, 2006 at 7:10 pm
Hi Colin. I was looking at the CXPACKETS originally. I noticed some very large values after re-indexing some dbs. I was trying to determine if hyper threading might be a problem...
August 18, 2006 at 11:35 am
I found some related links. This one describes similar results for the OLEDB counter.
http://www.sql-server-performance.com/forum/topic.asp?TOPIC_ID=8429&whichpage=1
This one has a bit of explaination of OLEDB, but not why the values appear...
August 18, 2006 at 10:35 am
It does rollover. It appears that an unsigned 4 byte integer is used for each counter. I've looked at the values of all counters on several machines. None are over 4.29E9. This...
August 18, 2006 at 10:25 am
Magarity, I'm curious. Can you tell us the actual performance improvement?
August 18, 2006 at 10:21 am
While I'm at it, here are the other wait types. I'm tempted to just ignore the OLEDB counter and remove it from "***Total***", but I am worried that something might...
August 17, 2006 at 11:29 am
Here are the results of logging since yesterday. I expect a SignalWaitTime rollover soon after about 4.3E9. The timing interval is 10 minutes. The jump from 29628864 to 3239083520 (37...
August 17, 2006 at 11:09 am
NOT IN is also bad. Perhaps trying a left join and testing for null in the where clause would be faster than using not exists. If there are a lot of columns...
August 16, 2006 at 6:45 pm
Viewing 15 posts - 106 through 120 (of 249 total)