|
|
Posted by dunleav1 on 11/28/18 11:58
I'm running my test locally so the execution time is the same whether
nocount is set or not. Thanks.
Erland Sommarskog wrote:
> dunleav1 (jmd@dunleavyenterprises.com) writes:
> > I can't use bulk load in my production code.
> > I'm creating test code to create load to tune my product.
> >
> > Here's the test code I have been running to tune i/o on mssql.
> > This is ran from tsql and not java. Running this code locally is about
> > 20% slower than loading Oracle via jdbc, but is a lot faster than
> > through jdbc.
> >
> > begin transaction
> > declare @c int
> > set @c = 0
> > while @c <500000
> > begin
> > set @c = @c + 1
> > insert into insertestjd calues (@c,@c,@c,@c)
> > end
> > commit transaction
>
> Do you run the above with SET NOCOUNT ON? If you don't, there is a
> (1 row affected) being sent to the client for every row inserted.
>
> When I added SET NOCOUNT ON, your script ran in 10 seconds from Mgmt
> Studio. I did not want to run it with SET NOCOUNT OFF.
>
> I also tried this:
>
> insert into inserttest
> select Number, Number, Number, Number
> from listtest..Numbers
> where Number < 500000
> go
>
> Where Numbers is a one-column with numbers from 1 to 999999 that I had
> around. This ran in six seconds.
>
>
> --
> Erland Sommarskog, SQL Server MVP, esquel@sommarskog.se
>
> Books Online for SQL Server 2005 at
> http://www.microsoft.com/technet/prodtechnol/sql/2005/downloads/books.mspx
> Books Online for SQL Server 2000 at
> http://www.microsoft.com/sql/prodinfo/previousversions/books.mspx
Navigation:
[Reply to this message]
|