Viewing 15 posts - 166 through 180 (of 211 total)
You will usually see checkpoints and log writes as suspended, these are run periodically by the system.
Is the blocked/blocking process in the output you've posted? I'm guessing it isn't.
June 12, 2011 at 11:18 am
Consider a calculated column:
create table t(
i int identity (1, 1),
data char(1),
url as 'http://www.blah.com/' + cast(i as varchar)
)
insert t (data) select 'a' union select 'b' union select 'c'
select *...
June 12, 2011 at 8:38 am
If we'd gone up to 50 our results may have been different, I'll grant you. But on a system with thousands of users there's plenty of inter-query parallelism going on...
June 10, 2011 at 9:31 pm
Yeah, fair point. Changing the threshold didn't work well for us in several cases (although admittedly we didn't go ridiculously high with it, mainly due to frustration), so we went...
June 10, 2011 at 9:55 am
In my opinion adding a unique clustered index on an identity column is quite reasonable. (But when doing so you just have to remember that you're doing it to "take...
June 10, 2011 at 9:43 am
One of many ways:
A package can have execute package tasks within the control flow. So if you have your "primary" package loading the data with the PK, it can call...
June 10, 2011 at 9:36 am
You could indeed use any of several existing (and sometimes free) out of the box tools. But you want to keep history (so there will necessarily be differences that you...
June 10, 2011 at 9:20 am
If all your index stats are up to date then I'd actually set it to 1, or at least find the offending statement and add the option maxdop 1 to...
June 10, 2011 at 9:09 am
I'm actually really bad with the GUI since I type stuff out 99% of the time, so I'm afraid I don't feel confident commenting on the steps you took. It...
June 10, 2011 at 9:01 am
I don't think it will make a noticeable difference. Either way the engine will run down the index to the required leaf to add data. If it's unique there isn't...
June 10, 2011 at 8:56 am
By the way, we were able to further improve the performance of this solution given some knowledge of business rules.
For instance, in an invoice table, posted invoices cannot be modified,...
June 10, 2011 at 12:08 am
You can use hash of the row data, transfer it across with the etl step, and use it as the basis for comparison.
In fact, I have code on my blog...
June 9, 2011 at 11:36 pm
tar (6/9/2011)
I still haven't had time to read any BOL on the subject but I will tomorrow.
I was looking for more of a performance information...
June 9, 2011 at 9:49 pm
Let me give you my understanding of what you have before I give my answer, so that if any of my assumptions are wrong you can correct them...
You have a...
June 9, 2011 at 7:32 pm
The index itself is still ordered (in a balanced tree structure) even if it is not the clustered index.
The special thing about a clustered index is that it is also...
June 9, 2011 at 6:34 pm
Viewing 15 posts - 166 through 180 (of 211 total)