|
Re: how data/log files grow in 2000/2005
4/02/2009 6:02:47 AM
(Total replies: 0)
|
| Rebuilding indexes (via DBCC or any other means) is fully logged & hence generates a lot of transaction logging. In total, this can require more space in the Tlog than the entire size of the .mdf b/c there is extra overhead associated with every TLog entry in addition to the row being logged.
So if you're going to rebuild indexes regularly, you need to ensure there is sufficient space in the log file, otherwise it has to grow during the rebuild phase, which adds substantial time to the ... |
|
|
|
|
|
|
Re: Finding the nth Row
27/02/2008 8:21:48 PM
(Total replies: 1)
|
| How can you expect to get the 5th row from a table with only 4 rows?
What do you expect to be returned if there's no row there? It might help to post a sample table with some rows (create table + insert statements) + expected resultset to clarify what you mean here.. |
|
|
|
Re: Finding the nth Row
28/02/2008 10:57:54 AM
(Total replies: 1)
|
| Assuming you can always use the literal value in the ClubID column, you can simply use WHERE for this, eg:
select top 1 *
from Club
where ClubID > 4
order by ClubID
If you need to get the n'th row, where there might be gaps in the ClubID ranges, use row_number() with a CTE, though this technique is less efficient, eg:
with orderedclubs as
(select row_number() over (order by clubid) as rownumber, * from club)
select clubid, name, email
from orderedclubs
where rownumber > ... |
|
|
|
|
|
|
Slide deck and discussion re July 07 meeting - Understanding Execution Plans (Greg Linwood)
17/07/2007 10:31:38 PM
(Total replies: 1)
|
| A copy of the slide deck for tonight's Melbourne SQL User Group presentation has been uploaded here:
http://www.sqlserver.org.au/meetings/Melb/2007/
If anyone has any questions or issues they'd like to discuss about the presentation or this topic in general please post here & I'll monitor this thread + try to respond asap.
Regards,
Greg Linwood |
|
|
|
Re: NO LOCK
28/06/2007 3:50:58 PM
(Total replies: 0)
|
| NO LOCK is not strictly an optimiser hint as it doesn't affect the execution plan chosen by the optimiser. It just stops locks being taken during query execution.
Setting the transaction isolation level accomplishes the same result but at the transaction level rather than statement level so which is better for you depends on whether you need NOLOCK to apply to multiple statements or not.
Can't remember how to do this in ADO.Net sorry :c/
Cheers,
Greg Linwood
|
|
|
|
Re: SQL not releasing memory
14/06/2007 5:20:02 PM
(Total replies: 1)
|
| SQL Server is not supposed to release memory en mass - this is intentional behaviour.
Unlike desktop applications which expect to share memory resources with other apps running on a desktop, SQL Server is designed to use memory as if it is the only application running on the server (because it usually is).
SQL Server lifts table data off disk & into memory whilst executing queries. To save it having to do this over & over (disk IO is very slow), it "caches" this data so that future ... |
|
|
|
Re: Resource for Meeting 2007-04-17 CLR integration
19/04/2007 10:29:01 AM
(Total replies: 1)
|
| Hi Zero
I'm just waiting on the resources from Greg at the moment. I'll post them & circulate an email letting everyone know they're available as soon as he sends them through.
Cheers,
Greg Linwood
|
|
|
|
Implementing Partitioned Tables
22/03/2007 7:15:47 AM
(Total replies: 0)
|
|
Geoff Orr's presentation material has now been uploaded to the website:
http://www.sqlserver.org.au/meetings/Melb/2007/03_GeoffOrr_ImplementingTablePartitioning
Please feel free to ask qns or discuss the meeting in this thread..
Cheers,
Greg Linwood |
|