Huge Performance Difference When Running UDF In Workstation Vs Server
Dec 13, 2007
Hi,
I created a CLR UDF that returns a large number of rows, when I run it from my VPC (XP, SQL Server Developer Edition and 1GB Memory) it takes approx 2 min and 30 secs to start displaying the rows (Using Management Studio), when I run the same query in our development server (Win 2003, SQL Server Enterprise Edition, 8 GB Memory and 8 Processors) it takes more than 15 min to start displaying the results, does anybody have an idea why is this happening?
I have encountered a problem with a specific set of tables. The same select yields slightly differing execution plans in two different environments (instances). But the slight variation seems to contain a huge differences in stats. I don't know the significance of these stats. The two tables have the exact same indices.
This is the selcet statement:
SELECT 'xx' FROM DUKS.dbo.Profiler WHERE DNA_Løbenummer IN (SELECT DNA_Løbenummer FROM DUKS.dbo.Effektregister WHERE Sagsnummer = '2015-00002')
I set up the collector, and specify the Run As as my AD account in the Collector Set - Properties - General screen. My AD account is the local admin of the remote server.
However, the collector does not seem to work. Although the collecting set is shown as running, the The blg file stays at 64K. If I open it, there is nothing inside (no counter at the bottom). What did I miss?
I'm having some problems debugging SQL Server stored procedures on a SQL Server 2005 server. I have installed Visual Studio 2005 on a workstation running Windows XP, now I'm trying to debug a ASP.Net web application that has some code that executes the stored procedures on a Windows 2003 Server running SQL Server 2005.
I opened VS2005 ... created a connection to the SQL Server 2005 instance ... open the Stored procedure ... right click the stored procedure name and selected Step into Stored Procedure and the following message is displayed:
Unable to start T-SQL debugging.Could not attach to SQL Server process on 'ServerName'.
I have a Stored Procedure (SP) that creates the data required for areport that I show on a web page. The SP does all the work and justreturns back a results set that I dump in an ASP.NET DataGrid. The SPtakes a product area and a start and end date as parameters.Here are the basics of the SP.1.Create temp table to store report results, all columns are createdthat will be needed at this point.2.Select products and general product data into the temp table.3.Create a cursor that loops through all the products in the temptable, running a more complex query with each individual product.4.The results of that query are updated on the temp table based on thecurrent product of the cursor.5.A complex "totals" query is run and the results from that areinserted into the temp table as the last 3 rows.In all we are talking about 120 rows in the temp table with 8 columnsthat are mostly numbers.I originally wrote this report SP about a month ago and it worked fine,ran in about 10 - 20 seconds based on server traffic and amount ofdata in the temp table. For the example I'm running there are the120 products.Just yesterday the (SP started timing out and when I ran the SPmanually from Query Analyzer (QA) (exec SP_NAME ... ) with the sameparameters as it was getting in the code it took 6 minutes to complete.I was floored. I immediately copied the SQL out of the SP and pastedinto another QA window, changed the variables to be hard coded valuesand ran it. It completed in 10 seconds.I'm really confused now. I ran a Profiler on the 2 when I ran themagain. The SQL code in QA executed again in ~10 seconds with 65,000reads. When the SP finished some 6 minutes later it had completed witthe right results but it needed 150,000,000 reads to do its job.How can the exact same SQL code produce such different results (time,disk reads) based on whether its in a SP or just run from QA but stillgive me the exact same output. The reports both look correct and havethe same numbers of rows.I asked my Sys Admin if he had done anything to anything and he saidno.I've been reading about recompiles and temp table indexes and allkinds of other stuff that could possibly be affecting it but havegotten nowhere.Any ideas are appreciated.
There is a stored procedure. It uses linked server. As we will be migrating to amazon cloud, our architect instructed not to replace linked server with openquery.
I'm having an issue with a query I'm running on Sql Server 2005. It's a semi-complex query involving an in-line table function and several left outer joins which are joined on to the results of the function call. Two of the left outer joins are then qualified in a where clause of the form where table.Col is not null; the idea is that the final result set contains data that has no match in those two tables.
The problem revolves around a where clause in the function and the last left outer join (ie, one of the ones qualified with where not null). When I alter the where clause of the function to further restrict the result set the function returns, the query times shoots up from 1 second to roughly 2-3 minutes. Note that the time the function takes to complete is not affected. The difference in time is purely down to what the query does with the results the function provides. Also note that the change to the where clause provides a subset of the original data; it does not add any more data (it actually restricts the original resultset by roughly 1000 rows).
I can bring the query speed back down again by removing the last left outer join - this join takes one of the columns from the function, and joins it to a small table - 924 rows. So it appears that this particular join is the cause of the issue, but only when using the resultset generated from the modified function query.
Now, as the thread title alludes, Sql Server 2000 and 2005 handle this differently, or appear to. When I execute this same query on a Sql 2000 machine, there's no apparent time differences, and the data that is returned is as expected. Does anyone have any suggestions as to what might be causing this and how I can fix it? I could simply return the larger resultset and use managed code to filter out the rows I don't want; however, I would like to get to the bottom of this, especially if it's going to effect future queries.
INSERT INTO MAIN VALUES ('1000', '1/1/2014',3000,1000,700,1500) INSERT INTO MAIN VALUES ('1000', '3/5/2014',1000,2000,650,200) INSERT INTO MAIN VALUES ('1000', '5/10/2014',500,5000,375,125) INSERT INTO MAIN VALUES ('1000', '11/20/2014',100,2000,400,300) INSERT INTO MAIN VALUES ('1000', '8/20/2014',100,3500,675,1300)
I am trying to write a query to calculate the running difference between data on different dates. Below is what my table of data looks like. Basically I want to calculate the difference between the total_completed for each state and date.
below is my code (I almost have what I need) I just can't figure out how show 0 as the completed_difference for the first Date for each state since there is no prior date to calculate against.
MRR_TOTALS_WEEK_OVER_WEEK AS ( SELECT T1.[Date] ,T1.States ,T2.Total_Completed ,ROW_NUMBER() OVER(PARTITION BY T1.States ORDER BY T1.States,T1.[Date]) AS ORDERING FROM TOTAL_CHARTS T1 LEFT JOIN TOTAL_COMPLETED T2 ON T1.[Date] = T2.[Date] AND T1.States = T2.States )
I have a medical DB with the loads 150,000 transactions per month. Each month, I load the tranactions into a table for the current year. I also have to update records for prior months based on current month information.
For example, out of 186,000 dump records...150,000 will be loaded into the main table and 36,000 will be used to update records already loaded into the main table.
The tables have 90 columns, I have a clustered PK using [Soc_Sec_Number] & [Month] & [Row Index]. I need the row index counter (like auto number in MS Access) because I can have multiple transactions per month for the same Soc Sec Number.
1) Load 150,000 records into main table (For december, this makes the table have 1,800,000 rows 2) Run queries for the remaining 36,000 rows to update records already loaded into the table containing the 1,800,000 rows. 3) The 36,000 queries have to be splits depending upon the update type code, So I am actually running 6 queries using 6,000 rows each against the 1,800,000 records.
The update queries are using inner joins with [Soc Sec #] and [Date], part of my composite Primary Key on both tables.
============================================================= Problem
This process takes forever, about 4 hours per monthly update. As the months go out, the main table gets larger and the time increases. It took almost 24 hrs to get from January 2004 to June 2004.
I am running Sql Server on my PC, no seperate workstation. My PC has 2.8 GHZ with about 1 Gig in RAM. Could my PC specs be too low. I noticed that the task mamager shows sqlservr.exe using over 657,000 mb of RAM when running.
I also ran a simple Select MAX(Soc_Sec_Number) query that took over 5 minutes. This is way too long especially since Soc_sec_Number is part of the composite PK.
Could my queries actually take that long or are my pC specs too low. MY PC seemed to freeze after the JUNE update? Any help appreciated.
Hi Everyone,Apparently, I was being asked on a question, "Why don't we procure adesktop PC to run MS SQL Server 2000 rather than a buying a server?".From a Management point-of-view, buying a desktop PC is much cheaperthan a server. However, I just wanted to understand that is it aviable solution given the database size is something around 200 GB?Equipping with more memory, more storage and a more powerful CPU on adesktop PC could really taking up the role to support the DBMS?Besides this "sensitive" costing concerns, what will be othersdifference in running the SQL Server 2000 on the two differenthardware architecture? For example, IO rate, reliability, RAID-1support, performance, … etc.(Note: The operating system is Microsoft Windows 2000 EnterpriseEdition)Regards,Ambrose
I am using SSIS to transfer data between two SQL Servers (2000). There is no transformation involved as the source and destination table structure is same. Even then the package execution takes lot of time.
The data in the tables is of the order of 66000000 the we were required to kill the package execution after it took more than 24 hours. The CPU usage was more than 13000s and disk I/O was well above 330000000. I am new to the tit-bits of SIS. Can anyone please tell me the reason as to why the package has gone so resource hungry.
was trying to performance tune a query and came up on this weird issue(Weird to me atleast.. possibly because I am a newbie:-))).
When I run the query mentioned below, I get the results in 1 second, while the same query, if I declare a variable for the value I am comparing, it takes 40 seconds. I am on SQL Server 2000. Both queries are being run in the same DB on the same tables and the only difference between the two is that the date has been declared as a variable. I am able to consistantly reproduce this. Tried casting the variable, declaring it as varchar , replacing the date by getdate() at both places, but the moment I use the variable, the performance goes for a toss. Has anyone came across any similar issues? TIA Callista
Query 1 select ae.Col1, ae.Col2, ae.Col3, ae.Col4, ge.Col4, c.Col5, 0 from table1 ae WITH (INDEX=ind_Col3 NOLOCK), table2 ge with (nolock), table3 c with (nolock) where ae.Col1 = ge.Col1 AND ae.Col4 = c.Col1 AND ae.Col3 > '2007-10-25 14:18:12.380' and ae.Col6 is not null
Query 2 declare @MyVariable datetimeselect @MyVariable = '2007-10-25 14:18:12.380' select ae.Col1, ae.Col2, ae.Col3, ae.Col4, ge.Col4, c.Col5, 0 from table1 ae WITH (INDEX=ind_Col3 NOLOCK), table2 ge with (nolock), table3 c with (nolock) where ae.Col1 = ge.Col1 AND ae.Col4 = c.Col1 AND ae.Col3 > @MyVariable and ae.Col6 is not null
select * from sys.dm_os_performance_counters returns the object names prefixed with "SQLServer:" (e.g. SQLServer:Databases)
It was expected as in other editions also. Issue is that when we try to crate "SQLServer Performance Condition alert", object names in "Object" list comes without the prefix "SQLServer:" (e.g. Databases). Please see the attached snapshots.
Dear All,We have a database which contains many tables which have millions ofrecords. When We attach the database with MS SQL Server 2005 StandardEdition Server and run some queries (having joins, filters etc.) thenthey take very long time to execute while when We execute same querieson Enterprise Edition then they run 10 times faster than on standardedition.Our database does not use any features which are present in EnterpriseEdition and not present in Standard Edition. We want to know what arethe differences between Standard Edition and Enterprise Edition forperformance. Why should we go for Enterprise Edition when StandardEdition has all the features required.We are presently using evaluation versions of SQL Server 2005 Standardand Enterprise Editions.Thanks and regards,Nishant Sainihttp://www.simplyjava.com
Hello,I have been searching and reading a lots of information on the microsoft website about the different version of SQL server, but still can not make my decision.In term of performance, is there a real big difference between the workgroup and the standard version? The workgroup is limited to 3 GB of RAM while the standard is unlimited, would that really change the performance if my server has 16Gb or RAM?The price difference is pretty substantial so if could only have to buy the workgroup , it would be better.One more question, regarding the type of licence, my server has 2 processors, could I avoid buying 2 licences and get the Server plus CAL instead. I am using this server to host 4 web-application running on SQL server. Each database is about 15 MB.Thanks in adavance for your advises.Arno
I would like to know if there is any meaningful difference in speed performance between using the DNS ("sql.server.com") or the IP address of the sql-server in the connection string. The advantage of using DNS is that if there is any change in IP, I do not have to change the connection strings, but I do not want to loose speed because of the necessity to resolve the DNS.
I have a table that will be loaded over night everyday and I need to write a query on running value difference ?
List of Columns (ID, Branch ,Group, Date, Value)
ID   Branch  Group  Date                 Value 1       A          C    2015-06-01           10 2       A          C      2015-06-02           15 3       A          C      2015-06-03           25 4        A          C      2015-06-04           20 5        B          D       2015-06-01           20 6        B          D       2015-06-02           25 7        B          D       2015-06-03           10 8        B          D       2015-06-04           20
I want the Output like below with a Running value difference in comparison to previous day.
ID   Branch  Group  Date         Value   Running Value 1       A          C    2015-06-01           10        10 2       A          C      2015-06-02           15         05 3       A          C      2015-06-03           25        10 4        A          C      2015-06-04           20        -5 5        B          D       2015-06-01           20        20 6        B          D       2015-06-02           25        05 7        B          D       2015-06-03           10        -15 8        B          D       2015-06-04           20        10
Basically I need to compare the previous day and show the difference. How can I do this in SQL 2008 r2?
I have one question what is performance difference between cluster index on numeric field or string field? I know that numeric is faster but why it is faster?
I have also posted this in microsoft.public.sqlserver.programming.
I have a query which, depending on where I run it from, will either take 10 milliseconds or 10 seconds.
The query works perfectly when run in SQL Server Management Studio... in my database of around 70,000 items it returns the results in around 10ms. It uses all my indexes and indexed views correctly.
However when I run the identical query from my ASP.NET application, it takes around 10 seconds... 1000 times longer. Looking at it in Sql Server Profiler I can't see any difference in the query, except from ASP.NET it needs 62531 reads and from SSMS it needs only 318 reads. If I copy the slow running ASP.NET query from the profiler into SSMS, then it runs quick again. The results returned are the same.
I have provided more details of the query below, but I guess my real question is: What is the best way to debug this? I'm not an expert with SQL Server, so any pointers on where I should start looking to find the difference in how the query is being executed would be a great help.
The query is of the form:
WITH RowPost AS ( SELECT ROW_NUMBER() OVER(ORDER BY DateCreated DESC) AS Row, ItemId, Title, .... FROM Items_View WITH(NOEXPAND) WHERE ItemX >= @minX AND ItemX <= @maxX AND ItemY >= @minY AND ItemY <= @maxY ) SELECT *, (SELECT Count(*) FROM RowPost) AS [Count] FROM RowPost WHERE Row >= @minRow AND Row < @maxRow
Where Items_View is an indexed view, and WITH(NOEXPAND) is being used to force it to use the indexed view (this is optimal). The line beginning "SELECT Count(*)" is to get the total number of results (without having to run the inner query a second time).
This is running against SQL Server Developer Edition.
SQL 2000 Connection String:user id=MyUserName;password=MyPassword;initial catalog=MyDB;server=MyServer;Connect Timeout=30 This SELECT statement returns its 10 results nearly instantly:SELECT * FROM MyTableDitto from above, but completes in 30-40 seconds:SELECT * FROM [dbo].[MyTable]Ditto from above, but completes nearly instantly: SELECT TOP 1000 * FROM [dbo].[MyTable] Obviously I have stopped using the [dbo] syntax in my SqlCommand's (SELECT's and EXECUTE's) but still would like to know why this is.vr, Rich
Hello, all, I started out thinking my problems were elsewhere but as Ihave worked through this I have isolated my problem, currently, as adifference between MSDE and SQL Express 2005 (I'll just call itExpress for simplicity).I have, to try to simplify things, put the exact same DB on twosystems, one running MSDE and one running Express. Both have 2 Ghzprocessors (one Intel, one AMD), both have a decent amount of RAM(Intel system has 1 GB, AMD system has 512 MB), and plenty of GB offree disk space. MSDE is running on the Intel system, Express isrunning on the AMD system. To keep things fair I use the exact sameDB's and query on both systems. The DB's were created on MSDE so Isp_detach_db'd them from MSDE and then sp_attach_db'd them to Express(this is how MS says to do a "side-by-side" upgrade, so it'sacceptable to do so). After fighting problems in performancedifferences in different situations I have narrowed the problem downto this:Executing a simple select statement with join clause on the databasesyields a difference in execution time that is quite great. Using theExpress Management program I can run the query against either system(MSDE or Express, the two systems are connected via crossover cable toeliminate any network problems/issues). When running the queryagainst the MSDE system (which is over the network) I consistently get<20 ms response times on the query. When running the query againstthe Express installation (which is in shared memory) I consistentlyget 700 ms or longer response times. Both times are for the TotalExecution Time.The query is simply this: select db1.* from db1.owner.tablename as db1inner join db2.owner.tablename as db2 on db1.pkey = db2.someid wheredb1.criteria = 3So, gimme all the columns from one table in one DB (local to theinstallation), matching the records in another DB (also local to theinstallation), where one field in the first db matches a field in thesecond db and where, in the first db, one column value = 3.The first table has a total record count of 630 records of which only12 match the where clause. The second table has a total record countof about 2,700 of which only 12 match up on the 12 out of 630.Even though the data is the same and I've done the detach and attach,and even done the sp_updatestats, the difference in execution time isremarkable, in a bad way.Checking the Execution Plan reveals that both queries have the samesteps, but, on the MSDE system the largest consumer in the process isthe Clustered Index Scan of the 630 record table (DB1 in my queryexample), using 85%. The next big consumer is a Clustered Index Seekagainst the other table (2,700 rows), using 15%.The Execution Plan against the Express system reveals basically theexact opposite: 27% going to the Clustered Index Scan of the 630record DB1, and 72% going to the Clustered Index Seek of the 2,700record DB2.I'm sorry to be stupid but I have this information but I don't knowwhat to do with it. The best that I can tell from this is that thisis the source of my problems. My problems are that on my currentsystems that my clients use the data is returned to them faster thanthey can click the mouse and that the new system (that is, when theychose (or are forced by attrition) to move to Vista and thus Express2005) the screen pop is like 1.5 seconds. This creates poor userexperience. Worse, one process I allow the users to do goes fromtaking 14-30 seconds to over 4 minutes (all on the same machine withthe same OS and version of my program, so it's not a machine or OS ormy app problem).Anyway, I hope someone can shed some light on this now that I've paredit down some.Thanks in advance.--HC
Executing the stored procedure took 45 seconds. But copying the code to a query window and setting up the variables (instead of parameters), it took 7 seconds.
In the query window, most of the processing cost (86%) is right up front in a "Distinct Sort." But in exec stored procedure, the cost for this step is 11% and the significant costs are in later "Table Scans."
I don't know why SQL Server would choose different execution plans when the code is identical in each.
I recently installed VB6.0 Enterprise on my system. As a part of the installation is the option to install SQL Server 6.5 Development. I elected to install SQL. The install went without any glitches. Upon going into SQL Enterprise Manager, I was asked to register the server where SQL was running. My understanding was that SQL would allow me to register the name of my host system. (This is a stand alone system with VB6.0 and Sql development 6.5 installed - no network attached). SQL will not allow me to register my server's name as the host server. It says that if cannot connect to the named server. SQLSERVER and SQL EXEC services appear to be running without any problems. Anyone have any idea where I should start or what could be disallowing my connection/registeration of my server. Any help would be greatly appreciated..... Thanks... Tom
Developers wanting to access sql server developer edition on XP workstation in small office network (workgroup) will have problems using windows authentication (trusted connection) if the XP "server" machine has "simple file sharing" enabled.
Simple file sharing in XP can be disabled in control panel - folder options.
For this sneak mode to work accounts with the same user and password should exist on both "client" and "server".
I have the following connection string in my Windows app that works fine on the Server: workstation id=SCSXP;packet size=4096;integrated security=SSPI;data source=scsxp;persist security info=False;initial catalog=nala But when I use that string on a workstation, it times out. I do have a Network connection to the Server. How should I modify the connection string to work on the workstation? Thanks, Bill
I am running SQL 7 (SP3) on a 2000 Server. When I configure the server as part of the Workgroup I have no problems connecting from the workstation. However, if I join the server to our 2000 domain, I cannot connect and/or register the server from the workstation through EM.
Is this a security issue? If so, what are the minimum requirements or what else am I missing. Any suggestions would be appreciated.
Could someone please tell me what I have to set at the Server side so that I will be able to STOP and START the SQL Server using Enterprise Manager and `sa` login remotely, say from the workstation.
This might not be a SQL Server question, might be something else. Ihave a workstation on a LAN that can't connect to the SQL Server. WhenI try to configure a User DSN to connect to it, the server name I wantappears TWICE in the list of possible servers. No matter which one Iselect, it can't get the connection.The workstation also can't connect through EM or QA.We've checked the user's permissions, we've re-installed some serverpacks, re-installed the Client Tools, we've even hunted through theregistry, everything looks OK, but we still can't connect. This is theonly machine on the LAN that can't connect. Everybody else connectsfine.I'm stumped. Any ideas? Anybody ever seen the server name appeartwice like that?-Emily
