How to improve a aquery

Question

Post reply

How to improve a aquery

Nubiam.stewart

Valued Member

Points: 70
More actions
September 26, 2013 at 3:49 am

#395560

Good morning,
I am writing a query thaht bring data from 7 tables.
In the first statge the data is stored in table variables.
Then I use 7 join to produce the data that I want to see.
The query takes 1 minute and 50 second to brin 22 records.
Could you please advise me how I can improve the performance of this query?
Thanks a lot,
Nubia

Viewing 11 posts - 1 through 10 (of 10 total)

You must be logged in to reply to this topic. Login to reply

Mr. Kapsicum SSCertifiable Points: 6128 More actions · Answer 1

need table structure and some sample data.

and Most. Imp. Your Query...!!! 🙂

BrainDonor SSCoach Points: 19254 More actions · Answer 2

The best way to get help with this is to include the code in question along with the definitions of the tables and some test data. Without those we're just guessing.

Have a look at the following article on how to post data that will help us to help you:

http://www.sqlservercentral.com/articles/SQLServerCentral/66909/

Steve Hall
Linkedin
Blog Site

Nubiam.stewart Valued Member Points: 70 More actions · Answer 3

Hello guys,

Thanks you for your reply. This is the definition of the sp. The data is taken from a ingress db. Unfotonately I don't have acces to the tables.

CREATE PROCEDURE [dbo].[usp_RPT_HSG_PropertyRepairsHistoryTSL]

(

@PropertyReferenceNumberNVARCHAR (max),

@StartDateDATETIME,

@EndDateDATETIME,

@GetCHRepTypeNVARCHAR (20)

)

AS

BEGIN

SET NOCOUNT ON

-- ===================================================

-- Execute no-lock expression at named linked servers

-- ===================================================

EXECUTE ('set lockmode session where readlock=nolock') AT INGRES2

DECLARE @PropertyRepairsHistory Table

(

JobnumberINT

,PropertyReferenceNVARCHAR (max)

,JobTitlenvarchar (60)

,Prioritynvarchar (20)

,JobTypenvarchar (20)

,Contractorcodenvarchar (4)

,Contractornvarchar (30)

,BilledValueMONEY

,ValuePaid MONEY

,StageHistDateDATE

,JobType2NVARCHAR (20)

)

INSERT INTO @PropertyRepairsHistory

Select * FROM OPENQUERY (ingres2, '

SELECT repheader.repairno

, repheader.propref

, repheader.repairdesc

, priority.decode

, reptype.code

, workforce.wforcecode

, workforce.wforcename

, worderline.billval

, worderline.payval

, statehist.statedate

, reptype.decode AS JobType

FROM((repheader INNER JOIN workorder ONrepheader.repairno = workorder.repairno) inner JOIN worderlineON workorder.repairno = worderline.repairno AND workorder.repsubno = worderline.repsubno)

LEFT OUTER JOIN reptype ON reptype.code = repheader.reptypenow

LEFT OUTER JOIN priorityONpriority.code = workorder.prioritycode

LEFT OUTER JOIN statehistON statehist.repaltkey = repheader.repairno

LEFT OUTER JOIN workforceON repheader.wforcecode = workforce.wforcecode

WHERE statehist.repstatecode = 99

')

DECLARE @Property Table

(

PropertyRefnvarchar(max)

,HouseNoNVARCHAR (4)

,AddressLine1nvarchar (30)

,AddressLine2nvarchar (30)

,AddressLine3nvarchar (30)

,Postcodenvarchar (8)

,Officenamenvarchar (max)

,Officecodenvarchar (max)

)

INSERT INTO @Property

Select * FROM OPENQUERY (ingres2, '

SELECT

propfixed.propref

,propfixed.houseno

, propfixed.proadd1

, propfixed.proadd2

, propfixed.proadd3

, propfixed.propstcde

,mnode.mnodename

, propfixed.mnodecode

FROM propfixedinner join mnode ON propfixed.mnodecode = mnode.code

' )

-- =======================

-- OutPut for Detail

-- =======================

SELECT *

FROM @PropertyRepairsHistory r inner join @Property P on R.PropertyReference = P.PropertyRef

WHERE r.StageHistDate BETWEEN @StartDate AND @EndDate

AND r.JobType IN (SELECT * FROM fn_CSVToTable(@GetCHRepType))

AND p.PropertyRef = @PropertyReferenceNumber

ORDER BY Jobnumber DESC

Thanks,

Nubia

Nubiam.stewart Valued Member Points: 70 More actions · Answer 4

Hi Briandonor,

I attach you the Estimated Execution plan and the number of rows per table.

According to the execution plan the more expensive statement is:

INSERT INTO @Property

Select * FROM OPENQUERY (ingres2, '

SELECT

propfixed.propref

,propfixed.houseno

, propfixed.proadd1

, propfixed.proadd2

, propfixed.proadd3

, propfixed.propstcde

,mnode.mnodename

, propfixed.mnodecode

FROM propfixedinner join mnode ON propfixed.mnodecode = mnode.code

I don't understand why if the statement is using a join table to bring data froma look up table that only has 20 records. the propfixed table has only 88923 records. Is there any other way to bring data from a look up table?

Thanks a lot,

Nubia

ccavaco Old Hand Points: 302 More actions · Answer 5

I believe you can get rid of the table variables altogether. Add the inner join used to populate the second table variable to the beginning of the first query to populate the first table variable and then add the additional "WHERE" clauses in the final query.

ChrisM@Work SSC Guru Points: 186127 More actions · Answer 6

How long do the two remote queries take to run and how many rows are returned?

SELECT *

FROM OPENQUERY (ingres2, '

SELECT repheader.repairno

, repheader.propref

, repheader.repairdesc

, priority.decode

, reptype.code

, workforce.wforcecode

, workforce.wforcename

, worderline.billval

, worderline.payval

, statehist.statedate

, reptype.decode AS JobType

FROM ((repheader

INNER JOIN workorder ON repheader.repairno = workorder.repairno)

inner JOIN worderline ON workorder.repairno = worderline.repairno

AND workorder.repsubno = worderline.repsubno)

LEFT OUTER JOIN reptype ON reptype.code = repheader.reptypenow

LEFT OUTER JOIN priority ON priority.code = workorder.prioritycode

LEFT OUTER JOIN statehist ON statehist.repaltkey = repheader.repairno

LEFT OUTER JOIN workforce ON repheader.wforcecode = workforce.wforcecode

WHERE statehist.repstatecode = 99')

SELECT *

FROM OPENQUERY (ingres2, '

SELECT

propfixed.propref

, propfixed.houseno

, propfixed.proadd1

, propfixed.proadd2

, propfixed.proadd3

, propfixed.propstcde

, mnode.mnodename

, propfixed.mnodecode

FROM propfixed

inner join mnode ON propfixed.mnodecode = mnode.code' )

At least one filter from your output detail section could be applied to the remote queries.

^{“Write the query the simplest way. If through testing it becomes clear that the performance is inadequate, consider alternative query forms.” - Gail Shaw}

For fast, accurate and documented assistance in answering your questions, please read this article.
Understanding and using APPLY, (I) and (II) Paul White
Hidden RBAR: Triangular Joins / The "Numbers" or "Tally" Table: What it is and how it replaces a loop Jeff Moden

Grant Fritchey SSC Guru Points: 398890 More actions · Answer 7

You're pumping 20000 rows (estimated, actual could be more) into table variables, where there are no statistics. Then you're combining that with a multi-statement table valued function, which also doesn't have statistics, to filter down to 20 rows. I'm surprised it's running as fast as it actually is.

You need to filter the data sooner as was suggested to you. Loading the data into table variables is probably unnecessary, but if you must do it that way, you're better off using temporary tables so that statistics can be generated to arrive a better execution plan.

"The credit belongs to the man who is actually in the arena, whose face is marred by dust and sweat and blood"
- Theodore Roosevelt

Author of:
SQL Server Execution Plans
SQL Server Query Performance Tuning

BrainDonor SSCoach Points: 19254 More actions · Answer 8

Thanks for that Grant. I saw the Table Variables but it just didn't register. "Looked But Failed To See" - it will be on my headstone one day.

Steve Hall
Linkedin
Blog Site

ChrisM@Work SSC Guru Points: 186127 More actions · Answer 9

Are the NVARCHAR(MAX) column types necessary?

DECLARE @PropertyRepairsHistory Table

(

Jobnumber INT

, PropertyReference NVARCHAR (max) --

, JobTitle nvarchar (60)

, Priority nvarchar (20)

, JobType nvarchar (20)

, Contractorcode nvarchar (4)

, Contractor nvarchar (30)

, BilledValue MONEY

, ValuePaid MONEY

, StageHistDate DATE

, JobType2 NVARCHAR (20)

)

DECLARE @Property Table

(

PropertyRef nvarchar(max) --

, HouseNo NVARCHAR (4)

, AddressLine1 nvarchar (30)

, AddressLine2 nvarchar (30)

, AddressLine3 nvarchar (30)

, Postcode nvarchar (8)

, Officename nvarchar (max) --

, Officecode nvarchar (max) --

)

^{“Write the query the simplest way. If through testing it becomes clear that the performance is inadequate, consider alternative query forms.” - Gail Shaw}

For fast, accurate and documented assistance in answering your questions, please read this article.
Understanding and using APPLY, (I) and (II) Paul White
Hidden RBAR: Triangular Joins / The "Numbers" or "Tally" Table: What it is and how it replaces a loop Jeff Moden

Nubiam.stewart Valued Member Points: 70 More actions · Answer 10

Good afternoon,

Thank you very much for your help.

I found some problems in the queries and lots of duplicates. I managed to spped up the query from 1 min. 30 seconds to 30 seconds using a different aproach.

Cheers,

Nubia:-)