February 20, 2013 at 10:28 pm
I have a extremely time consuming query that times out whenever it runs. To cure the issue, I created a view of the lengthy query then a table of the view. Next, I plan to add a maintenance plan that would Drop the view then create the view, then drop the table the create the table. Is this the most efficient method to achieve my goal of improve performance? The resulting query return large chunks of data very quickly versus timing out. I plan to post the plan some time in the morning.
Thanks.
February 21, 2013 at 2:07 am
I'd say that the first thing you should try is some query optimisation - maybe you can get the query running fast enough so that you don't need to do this.
Otherwise, I'd consider creating a physical table which has the query's structure and then building a stored proc which truncates/rebuilds this table using your query. Then you can call this stored proc as part of your Agent job & then use the table in what comes next ...
The absence of evidence is not evidence of absence
- Martin Rees
The absence of consumable DDL, sample data and desired results is, however, evidence of the absence of my response
- Phil Parkin
February 21, 2013 at 1:01 pm
GO
Drop view AssemblyDefects
GO
create View AssemblyDefects AS
(select p.model, i.defect_title, p.item, d.sn, d.dateinspected, t.type, i.defect_id
from productiondefect as d, assignworkorder as p, tlkp_defects as i, tlkp_item as t
where substring(d.sn, 7, 5) = p.workorder
and d.defect_id = i.defect_id
and p.item = t.item)
GO
DROP Table ADR
GO
select *
into ADR
from AssemblyDefects
February 21, 2013 at 1:17 pm
Rewrite the query or maybe creating a computed column on substring(dsn, 7, 5) and creating index on that column will help this query.
February 21, 2013 at 1:21 pm
Do you know why dropping and recreating makes things go faster, and they gradually slow down?
Do you have maintenance that updates statistics, and reindexes, if needed?
Plus, if the code within the view is indicative of your coding, there is likely a lot of optimization that can occur.
You are using the old style joins, which functionally are the same, but are not very readable.
Also, doing the join on the substring(sn, 7, 5) is referred to as NON-sargable. Please research this further.
I took the liberty of re-writing thie code. There are many other ways to do this, but since I am at the office, you are only getting the "quick" version.
select
p.model,
i.defect_title,
p.item,
d.sn,
d.dateinspected,
t.type,
i.defect_id
from
(SELECT sn,
substring(sn, 7, 5) as [sn_join]
dateinspected,
defect_id
FROM productiondefect) D
INNER JOIN assignworkorder p ON d.[sn_join] = p.workorder
INNER JOIN tlkp_defects i ON d.defect_id = i.defect_id
INNER JOIN tlkp_item as t ON p.item = t.item
Michael L John
If you assassinate a DBA, would you pull a trigger?
To properly post on a forum:
http://www.sqlservercentral.com/articles/61537/
Viewing 5 posts - 1 through 4 (of 4 total)
You must be logged in to reply to this topic. Login to reply