How to increase performance of this query? - sql-server

I have an SQL query, it is running on MSSQL 2008 R2
View vMobileLastMobileHistory has about 1000 rows and
select * from vMobileLastMobileHistory is taking 0.2 sec
but this query is taking 5 seconds, how can I optimize this code?
(I think the problem is INTERSECT but I dont know how change this)
SELECT DISTINCT *
FROM
(
SELECT vMobileLastMobileHistory.*
FROM vMobileLastMobileHistory
LEFT OUTER JOIN MobileType_DomainAction ON
MobileType_DomainAction.tiMobileType = vMobileLastMobileHistory.tiMobileType
LEFT OUTER JOIN MobileType_User ON
MobileType_User.MobileID = MobileType_DomainAction.ID
WHERE MobileType_User.UserID = #UserID OR #UserID = - 1
INTERSECT
SELECT vMobileLastMobileHistory.*
FROM vMobileLastMobileHistory
LEFT OUTER JOIN dbo.Region_User ON
dbo.vMobileLastMobileHistory.strRegion = dbo.Region_User.strRegion
WHERE Region_User.iSystemUser = #UserID OR #UserID = - 1
INTERSECT
SELECT vMobileLastMobileHistory.*
FROM vMobileLastMobileHistory
LEFT OUTER JOIN Contractor_User ON
vMobileLastMobileHistory.strContractor = Contractor_User.strContractor
WHERE Contractor_User.iSystemUser = #UserID OR #UserID = - 1
)

The problem is that if you have any indexes on your iSytemUser columns, the optimise is unable to use them because it has to account for a specific userID being passed, or returning all results, it would be better to logically separate your two cases. In addition, since you don't care about any columns in the auxiliary tables, you could use EXISTS in your case of specific users to take advantage of a semi join:
IF (#UserID = -1)
BEGIN
SELECT DISTINCT *
FROM vMobileLastMobileHistory;
END
ELSE
BEGIN
SELECT DISTINCT *
FROM vMobileLastMobileHistory AS mh
WHERE EXISTS
( SELECT 1
FROM Contractor_User AS cu
WHERE cu.strContractor = mh.strContractor
AND cu.iSystemUser = #UserID
)
AND EXISTS
( SELECT 1
FROM Region_User AS ru
WHERE ru.strRegion = mh.strRegion
AND ru.iSystemUser = #UserID
)
AND EXISTS
( SELECT 1
FROM MobileType_DomainAction AS da
INNER JOIN MobileType_User AS mu
ON mu.MobileID = da.ID
WHERE da.tiMobileType = mh.tiMobileType
AND mu.iSystemUser = #UserID
);
END
Now you can have two execution plans for each case (returning all results, or for a specific user), in each case you only need to read from vMobileLastMobileHistory once, and you also limit the sorts required by removing the INTERSECT and replacing with 3 EXISTS clauses.
If they don't already exist then you may also which to consider some indexes on your tables. A good way of finding out what indexes would help is to run the query in SQL Server Management Studio with the option "Show Actual Execution Plan" enabled, this will then show you any missing indexes.

Most of time Intersect and Inner Join will be same. You are not share your data, so based on my knowledge and this link, I just replace intersect query into Inner join query as :
--I think you don't need distinct upper query. If you have issue inform me.
SELECT DISTINCT vml.*
FROM vMobileLastMobileHistory vml
LEFT OUTER JOIN MobileType_DomainAction mtda ON mtda.tiMobileType = vml.tiMobileType
LEFT OUTER JOIN MobileType_User ON MobileType_User.MobileID = mtda.ID
LEFT OUTER JOIN dbo.Region_User ON dbo.vml.strRegion = dbo.Region_User.strRegion
LEFT OUTER JOIN Contractor_User ON vml.strContractor = Contractor_User.strContractor
WHERE
(MobileType_User.UserID = #UserID
and Region_User.iSystemUser = #UserID or Contractor_User.iSystemUser = #UserID
) OR #UserID = - 1

Related

Combine multiple SQL statements into one stored procedure

Okay from these two tables:
SELECT *
FROM Sessions
JOIN Sessions ON Sessions.ID = Sessions.SessionID
I need to count how many seats are taken (counting how many of each sessionID exist)
SELECT DISTINCT
Sessions.SessionID,
COUNT(*) OVER (PARTITION BY SessionID) AS Sess
FROM
Sessions
WHERE
UserID = #UserId
And then if the count Sess is less (an int column in table), select only those sessions from my join (below code is not correct):
SELECT *
FROM Sessions
WHERE UserId = #UserId
How do I combine all of these statements into one stored procedure?
Each select statement works on its own, but I don't know how to combine this into one solution.
WITH EligibleSessions As (
SELECT s.ID
FROM Sessions s
JOIN RegistrantSessions rs ON s.ID = rd.SessionID
WHERE rs.EventID = #EventID
GROUP BY s.ID, s.SeatLimit
HAVING COUNT(rs.*) < s.SeatLimit
)
SELECT s.*
FROM Sessions
INNER JOIN EligibleSessions es on es.ID = s.ID
WHERE s.EventId = #EventId
AND s.SessionTime = #SessionTime
AND s.SessionType = #SessionType
AND s.Active = #SessionActive
There seems like an extra layer of nesting here, but I didn't want to have to include every field in the Sessions table in the GROUP BY.
You may need to repeat some of the WHERE conditions in the CTE. It's also weird to me for the EventID field to repeated in both Sessions and RegistrantSessions. Seems like an indication something is not normalized properly.
If I understand your table structure correctly, we could join a the count query as subselect too:
SELECT *
FROM Sessions
JOIN (
SELECT DISTINCT RegistrantSessions.SessionID
,COUNT(*) OVER (PARTITION BY SessionID) AS SeatCount
FROM RegistrantSessions
JOIN RegistrantSessions ON Sessions.ID = RegistrantSessions.SessionID
WHERE EventID = #EventId
) AS regCounts ON regCounts ON Sessions.ID = regCounts.SessionID
WHERE Sessions.SeatLimit < regCounts.SeatCount
AND EventId = #EventId
AND SessionTime = #SessionTime
AND SessionType = #SessionType
AND Active = #SessionActive

Updating DISTINCTROW in SQL Server [duplicate]

What would the syntax be to convert this MS Access query to run in SQL Server as it doesn't have a DistinctRow keyword
UPDATE DISTINCTROW [MyTable]
INNER JOIN [AnotherTable] ON ([MyTable].J5BINB = [AnotherTable].GKBINB)
AND ([MyTable].J5BHNB = [AnotherTable].GKBHNB)
AND ([MyTable].J5BDCD = [AnotherTable].GKBDCD)
SET [AnotherTable].TessereCorso = [MyTable].[J5F7NR];
DISTINCTROW [MyTable] removes duplicate MyTable entries from the results. Example:
select distinctrow items
items.item_number, items.name
from items
join orders on orders.item_id = items.id;
In spite of the join getting you the same item_number and name multiple times when there is more than one order for it, DISTINCTROW reduces this to one row per item. So the whole join is merely for assuring that you only select items for which exist at least one order. You don't find DISTINCTROW in any other DBMS as far as I know. Probably because it is not needed. When checking for existence, we use EXISTS of course (or IN for that matter).
You are joining MyTable and AnotherTable and expect for some reason to get the same MyTable record multifold for one AnotherTable record, so you use DISTINCTROW to only get it once. Your query would (hopefully) fail if you got two different MyTable records for one AnotherTable record.
What the update does is:
update anothertable
set tesserecorso = (select top 1 j5f7nr from mytable where mytable.j5binb = anothertable.gkbinb and ...)
where exists (select * from mytable where mytable.j5binb = anothertable.gkbinb and ...)
But this uses about the same subquery twice. So we'd want to update from a query instead.
The easiest way to get one result record per <some columns> in a standard SQL query is to aggregate data:
select *
from anothertable a
join
(
select j5binb, j5bhnb, j5bdcd, max(j5f7nr) as j5f7nr
from mytable
group by j5binb, j5bhnb, j5bdcd
) m on m.j5binb = a.gkbinb and m.j5bhnb = a.gkbhnb and m.j5bdcd = a.gkbdcd;
How to write an updateble query is different from one DBMS to another. Here is the final update statement for SQL-Server:
update a
set a.tesserecorso = m.j5f7nr
from anothertable a
join
(
select j5binb, j5bhnb, j5bdcd, max(j5f7nr) as j5f7nr
from mytable
group by j5binb, j5bhnb, j5bdcd
) m on m.j5binb = a.gkbinb and m.j5bhnb = a.gkbhnb and m.j5bdcd = a.gkbdcd;
The DISTINCTROW predicate in MS Access SQL removes duplicates across all fields of a table in join statements and not just the selected fields of query (which DISTINCT in practically all SQL dialects do). So consider selecting all fields in a derived table with DISTINCT predicate:
UPDATE [AnotherTable]
SET [AnotherTable].TessereCorso = main.[J5F7NR]
FROM
(SELECT DISTINCT m.* FROM [MyTable] m) As main
INNER JOIN [AnotherTable]
ON (main.J5BINB = [AnotherTable].GKBINB)
AND (main.J5BHNB = [AnotherTable].GKBHNB)
AND (main.J5BDCD = [AnotherTable].GKBDCD)
Another variant of the query.. (Too lazy to get the original tables).
But like the query above updates 35 rows =, so does this one
UPDATE [Albi-Anagrafe-Associati]
SET
[Albi-Anagrafe-Associati].CRegDitte = [055- Registri ditte].[CRegDitte],
[Albi-Anagrafe-Associati].NIscrTribunale = [055- Registri ditte].[NIscrTribunale],
[Albi-Anagrafe-Associati].NRegImprese = [055- Registri ditte].[NRegImprese]
FROM [055- Registri ditte]
WHERE EXISTS(
SELECT *
FROM [055- Registri ditte]-- [Albi-Anagrafe-Associati]
WHERE ([055- Registri ditte].GIBINB = [Albi-Anagrafe-Associati].GKBINB)
AND ([055- Registri ditte].GIBHNB = [Albi-Anagrafe-Associati].GKBHNB)
AND ([055- Registri ditte].GIBDCD = [Albi-Anagrafe-Associati].GKBDCD))
Update [AnotherTable]
Set [AnotherTable].TessereCorso = MyTable.[J5F7NR]
From [AnotherTable]
Inner Join
(
Select Distinct [J5BINB],[5BHNB],[J5BDCD]
,(Select Top 1 [J5F7NR] From MyTable) as [J5F7NR]
,[J5BHNB]
From MyTable
)as MyTable
On (MyTable.J5BINB = [AnotherTable].GKBINB)
AND (MyTable.J5BHNB = [AnotherTable].GKBHNB)
AND (MyTable.J5BDCD = [AnotherTable].GKBDCD)

SQL Query to get data from various databases

I wrote the below query to pull the data from different databases. I have created two temp tables to pull the data from two different databases and finally a select statement from the original database to join all the tables. My query is getting executed but not getting any data.(Report is blank). I tried executing the two temp tables separately. it is giving the correct data. But when I execute the whole query, the result is blank. Below is the query. Please help.
"set fmtonly off
use GODSDB
IF object_id('tempdb..#CISIS_Call_Log') IS NOT NULL DROP TABLE #CISIS_Call_Log
select *
into #CISIS_Call_Log
from OPENQUERY (CSISDB,
'select
ccl.ContractOID,
ccl.db_insertdate,
ccl.ContractCallLogStatusIdentifier,
ccl.db_UpdateDate,
ccp.ContractCallLogPurposeOID,
ccp.ContractCallLogPurposeIdentifier,
ccp.Description
from csisdb.dbo.ContractCallLog CCL
inner join csisdb.dbo.ContractCallLogPurpose CCP on ccl.ContractCallLogPurposeIdentifier = ccp.ContractCallLogPurposeIdentifier
where JurisdictionShortIdentifier = ''ON''
AND ContractCallLogStatusIdentifier IN (''DNR'', ''NR'')
')
IF object_id('tempdb..#CMS_Campaign') IS NOT NULL DROP TABLE #CMS_Campaign
select *
into #CMS_Campaign
from OPENQUERY (BA_GBASSTOCMS, '
Select
SystemSourceIdentifier,
ContractOID,
OfferSentDate,
CampaignOfferTypeIdentifier,
CampaignContractStatusIdentifier,
CampaignContractStatusUpdateDate,
DeclineDate,
CampaignOfferOID,
CampaignOID,
CampaignStartDate,
CampaignEndDate,
Jurisdiction,
CampaignDescription
from CMS.dbo.vw_CampaignInfo
where Jurisdiction = ''ON''
and CampaignOfferTypeIdentifier = ''REN''
')
select mp.CommodityTypeIdentifier as Commodity
,c.RtlrContractIdentifier as ContractID
,cs.ContractStatusIdentifier as ContractStatus
,c.SigningDate
,cf.StartDate as FlowStartDate
,cf.EndDate as FlowEndDate
,datediff(day, getdate(), c.RenewalDate) as RemainingDays
,c.RenewalDate
,l.ContractCallLogStatusIdentifier as CallLogType
,Substring (l.Description, 1, 20) as CallPurpose
,l.db_insertDate as CallLogDate
,cms.CampaignOfferOID as OfferID
,cms.CampaignContractStatusIdentifier as OfferStatus
,cms.CampaignContractStatusUpdateDate as StatusChangeDate
,cms.DeclineDate
from Contract c
inner join contractstate cs on cs.contractoid = c.ContractOID
and cs.ContractStatusIdentifier in ('ERA', 'FLW')
and datediff(day, getdate(), c.RenewalDate) > 60
inner join SiteIdentification si on si.SiteOID = c.SiteOID
inner join MarketParticipant mp on mp.MarketParticipantOID = si.MarketParticipantOID
inner join Market m on m.MarketOID = mp.MarketOID
inner join Jurisdiction j on j.JurisdictionOID = m.JurisdictionOID
and j.CountryCode = 'CA'
and j.ProvinceOrStateCode = 'ON'
inner join ContractFlow cf on cf.ContractOID = c.ContractOID
inner join #CISIS_Call_Log l on convert(varchar(15), l.ContractOID) = c.RtlrContractIdentifier
inner join #CMS_Campaign cms on convert(varchar(15), cms.ContractOID) = c.RtlrContractIdentifier
set fmtonly on"
IF the data in each temp table is verified, then:
Try a smaller, less complex, query to test your temp tables with. Also try them using a LEFT join as well e.g.:
select
c.RtlrContractIdentifier as ContractID
, c.SigningDate
, datediff(day, getdate(), c.RenewalDate) as RemainingDays
, c.RenewalDate
, l.ContractCallLogStatusIdentifier as CallLogType
, Substring (l.Description, 1, 20) as CallPurpose
, l.db_insertDate as CallLogDate
, cms.CampaignOfferOID as OfferID
, cms.CampaignContractStatusIdentifier as OfferStatus
, cms.CampaignContractStatusUpdateDate as StatusChangeDate
, cms.DeclineDate
from Contract c
LEFT join #CISIS_Call_Log l on convert(varchar(15), l.ContractOID) = c.RtlrContractIdentifier
LEFT join #CMS_Campaign cms on convert(varchar(15), cms.ContractOID) = c.RtlrContractIdentifier
Does this return data? Does it return data from both joined tables?
If neither temp table is returning data then those join conditions need to be changed.
If both temp tables do return data from that query, then try INNER joins. If that still works, then add back more joins (one at a time) until you identify the join that causes the overall fault.
Without data for every table it just isn't possible for us to pinpoint the exact reason for a NULL result. Only you can, so you need to trouble-shoot the problem one step at a time.

SQL Server: breaking a multiple join query into smaller ones and creating a function

I have a complex query with multiple joins in it, which runs multiple times in my application. I want to write this query as a function by breaking this query into smaller pieces inside the function. As a newbie, I have limited knowledge on SQL Server.
The following is the query:
SELECT
ts.lable as label,
ts.percentage as rate
FROM
TaxSet ts
JOIN
UserInfo u ON u.userID = ?
AND u.countryID = ts.countryId
AND (ts.stateId IS NULL OR ts.stateId = 0 OR LEN(ts.stateId) < 1)
JOIN
Users us ON u.userID = us.id
JOIN
Users p ON us.parentID = p.id
AND ts.ispID = p.id
JOIN
ProductType pt ON ts.productTypeID = pt.id
WHERE
startDate <= getutcdate()
AND getutcdate() <= endDate
AND pt.identifier = ?
AND ts.id NOT IN (SELECT eu.ispTaxSettingId
FROM ExemptUsers eu
WHERE eu.ExemptUserId = ?)
Now, how can I write a function by breaking this query into smaller ones.
Thanks in advance.
May I ask why you want to split it into functions? I reformatted your code and have put it into a stored procedure for now. My thinking is that you want to pass through a Identifier and UserID which are the parameters of your query.
I have modified the query and removed the Not In statement. This has been replaced by a LEFT JOIN to ExemptUsers on u.id = eu.ExemptUserID and then an addition to the WHERE clause to ensure eu.ExemptUserID is NULL. This is basically a clearer way of saying "If the userID exists in table ExemptUsers do not bring back results for that user".
In addition I have removed the join to Users p as I can't see that this was being used in any way, unless you want to ensure that the user has a parent?
CREATE PROCEDURE wsp_StoredProcName
(#UserID int,
#Identifier int)
AS
BEGIN
SELECT
ts.lable as label,
ts.percentage as rate
FROM
TaxSet ts
INNER JOIN UserInfo u ON u.userID = ts.UserID
AND u.countryID = ts.countryId
INNER JOIN Users us on u.userID = us.id
INNER JOIN ProductType pt on ts.productTypeID = pt.id
LEFT JOIN ExemptUsers eu on u.id = eu.ExemptUserID
WHERE
(
ts.UserID = #UserID
and pt.identifier = #Identifier
and startDate <= getutcdate()
and getutcdate() <= endDate
AND eu.ExemptUserID IS NULL
and
(
ts.stateId is null or ts.stateId = 0 or len(ts.stateId) < 1
)
)
END
After all you select from TaxSet where certain conditions must be met: The date range, the state, a relation to a particular user and its parent, a relation to a particular product type, and the non-existence for a particular exempt user. So use EXISTS and NOT EXISTS throughout your query in order to make it plain to the reader and dbms what you want to achieve. The more straight-forward a query, the easier it often is for the optimizer to deal with it.
select
ts.lable as label,
ts.percentage as rate
from taxset ts
where getutcdate() between ts.startdate and ts.enddate
and (stateid is null or stateid between 0 and 9)
and exists
(
select *
from users u
join userinfo ui on ui.userid = u.id
where u.id = ?
and ui.countryid = ts.countryid
and u.parentid = ts.ispid
)
and exists
(
select *
from producttype pt
where pt.identifier = ?
and pt.id = ts.producttypeid
)
and not exists
(
select *
from exemptusers eu
where eu.exemptuserid = ?
and eu.isptaxsettingid = ts.id
);
As others have mentioned: When splitting a query into smaller parts and execute these separately, it usually becomes slower not faster. This is due to the fact that a dbms is made to do exactly this in the most efficient way internally. Of course sometimes a dbms' optimizer fails to find a good execution plan, though. You may want to look at the execution plan and check whether you find that plan appropriate.

SQL Server, table-valued functions slow processing

I tried everything but I couldn't overcome this problem.
I have a table-valued function.
When I call this function with
SELECT * FROM Ratings o1
CROSS APPLY dbo.FN_RatingSimilarity(50, 497664, 'Cosine') o2
WHERE o1.trackId = 497664
It takes a while to be executed. But when I do this.
SELECT * FROM Ratings o1
CROSS APPLY dbo.FN_RatingSimilarity(50, o1.trackId, 'Cosine') o2
WHERE o1.trackId = 497664
It is executed in 32 seconds. I created all indexes but It didn't help.
My function by the way:
ALTER FUNCTION [dbo].[FN_RatingSimilarity]
(
#trackId INT,
#nTrackId INT,
#measureType VARCHAR(100)
)
RETURNS TABLE
WITH SCHEMABINDING
AS
RETURN
(
SELECT o2.id,
o2.name,
o2.releaseDate,
o2.numberOfRatings,
o2.averageRating,
COUNT(1) as numberOfSharedUsers,
CASE #measureType
WHEN 'Cosine' THEN SUM(o3.score*o4.score)/(0.01+SQRT(SUM(POWER(o3.score,2))) * SQRT(SUM(POWER(o4.score,2))))
WHEN 'AdjustedCosine' THEN SUM((o3.score-o5.averageRating)*(o4.score-o5.averageRating))/(0.01+SQRT(SUM(POWER(o3.score-o5.averageRating, 2)))*SQRT(SUM(POWER(o4.score-o5.averageRating, 2))))
WHEN 'Pearson' THEN SUM((o3.score-o1.averageRating)*(o4.score-o2.averageRating))/(0.01+SQRT(SUM(POWER(o3.score-o1.averageRating, 2)))*SQRT(SUM(POWER(o4.score-o2.averageRating, 2))))
END as similarityRatio
FROM dbo.Tracks o1
INNER JOIN dbo.Tracks o2 ON o2.id != #trackId
INNER JOIN dbo.Ratings o3 ON o3.trackId = o1.id
INNER JOIN dbo.Ratings o4 ON o4.trackId = o2.id AND o4.userId = o3.userId
INNER JOIN dbo.Users o5 ON o5.id = o4.userId
WHERE o1.id = #trackId
AND o2.id = ISNULL(#nTrackId, o2.id)
GROUP BY o2.id,
o2.name,
o2.releaseDate,
o2.numberOfRatings,
o2.averageRating
)
Any help would be appreciated.
Thanks.
Emrah
I believe that your bottleneck is the calculations + your very expensive inner joins.
The way your are joining is basically creating a cross join - It is returning a result set with all ther records linked to all other records, Except the one for which the id is supplied. Then you go and add to that result set with the other inner joins.
For every inner join, SQL goes and creates a result set with all the rows matching.
So the first thing you do in your query is to tell SQL to basically do a cross join on the same table. (I am assuming you are still following, that looks pretty advanced so I'll just take you are familiar with advanced SQL syntax and operators)
Now in the next inner join, you are applying the Results table to your newly created huge result set, and only then filtering out the ones not both tables.
So as a start, see if you can't do your joins the other way around. (This really depends on your table record count and record sizes). Try to have the smallest result sets first and then join onto that.
The second thing you might want to try is to firstly limit your result set even before the joins.So start with a CTE where you filter for o1.id = #trackId. Then select * from this CTE , do your joins on the CTE and then filter in your query for o2.id = ISNULL(#nTrackId, o2.id)
I will work on an example, stay tuned...
--
Ok, I added an example, did a quick test and the values returned are the same. Run this through your data and let us know if there is any improvement. (Note, this does not address the INNER JOIN order point discussed, still do play around with that.)
Example:
ALTER FUNCTION [dbo].[FN_RatingSimilarity_NEW]
(
#trackId INT,
#nTrackId INT,
#measureType VARCHAR(100)
)
RETURNS TABLE
WITH SCHEMABINDING
AS
RETURN
(
WITH CTE_ALL AS
(
SELECT id,
name,
releaseDate,
numberOfRatings,
averageRating
FROM dbo.Tracks
WHERE id = #trackId
)
SELECT o2.id,
o2.name,
o2.releaseDate,
o2.numberOfRatings,
o2.averageRating,
COUNT(1) as numberOfSharedUsers,
CASE #measureType
WHEN 'Cosine' THEN SUM(o3.score*o4.score)/(0.01+SQRT(SUM(POWER(o3.score,2))) * SQRT(SUM(POWER(o4.score,2))))
WHEN 'AdjustedCosine' THEN SUM((o3.score-o5.averageRating)*(o4.score-o5.averageRating))/(0.01+SQRT(SUM(POWER(o3.score-o5.averageRating, 2)))*SQRT(SUM(POWER(o4.score-o5.averageRating, 2))))
WHEN 'Pearson' THEN SUM((o3.score-o1.averageRating)*(o4.score-o2.averageRating))/(0.01+SQRT(SUM(POWER(o3.score-o1.averageRating, 2)))*SQRT(SUM(POWER(o4.score-o2.averageRating, 2))))
END as similarityRatio
FROM CTE_ALL o1
INNER JOIN dbo.Tracks o2 ON o2.id != #trackId
INNER JOIN dbo.Ratings o3 ON o3.trackId = o1.id
INNER JOIN dbo.Ratings o4 ON o4.trackId = o2.id AND o4.userId = o3.userId
INNER JOIN dbo.Users o5 ON o5.id = o4.userId
WHERE o2.id = ISNULL(#nTrackId, o2.id)
GROUP BY o2.id,
o2.name,
o2.releaseDate,
o2.numberOfRatings,
o2.averageRating
)

Resources