Please start any new threads on our new
site at https://forums.sqlteam.com. We've got lots of great SQL Server
experts to answer whatever question you can come up with.
| Author |
Topic |
|
marko_1
Starting Member
1 Post |
 Posted - 2012-11-08 : 12:31:01
|
I was asked to think of an example database in which an index would improve the performance of a low hit rate query.
declare @prsi_no int
declare @L int, @i int, @x datetime, @y datetime, @z real, @u datetime
declare @s varchar(1000)
set @x = getdate() -- start time
select x = @x
set @i=1
while @i < 10000 -- N
begin
begin tran
set @s = 'select count(*) from dbo.Asg1_tbl where prsi_no = ' + str(@prsi_no)
exec(@s)
commit tran
set @i = @i + 1
end
set @y= getdate()
select y = @y
set @z = datediff(ms, @x, @y)
select 'Run time = ' + cast(@z as varchar(20)) + ' miliseconds.'
go
select * from dbo.Asg1_tbl where prsi_no = ceiling (rand()*10000)
I used this code to test the speed with a primary key on and with no primary key to test the speed improvement..
Anyone have an what code to use to generate a random number and also how to tabulate results?? |
|
|
chadmat
The Chadinator
1974 Posts |
Posted - 2012-11-09 : 03:17:41
|
@prsi_no is NULL for every execution. Is that your intention?
-Chad |
 |
|
|
|
|
|
Performance improvement on low hit rate query