I've implemented a UDF in SQL Server 2005 written in C#. The function with its assembly has been registered ok with SQL Server and works fine. It accepts three short strings (nvarchar of lengths 5, 35, and 35) and returns a SQL formatted string (SqlString).
When I run the function to test it it works just fine, and the same is true if I run the function inside a cursor to update a field in a table. But when I do a simple update it crashes. I've so far received two different errors: first one error saying a string could not be converted into an integer (but the error does not occur when I enter the same input values manually via a test Windows form, or through the new Query Analyzer as a single query - or using it inside a cursor). Then one error saying a string was too short (I couldn't use substring(X, Y) because the string, I was told, was too short - it wasn't).
The problem thus cannot be with the function since it works just fine if I do like this:
UPDATE myTable SET CodeField = dbo.fnMyFunction(Field1, Field2, Field3) WHERE PersonId = 10000001
And it works fine while doing the same thing inside a cursor (for instance working with the first 10, 100 or 1000 records).
But when I do this it crashes:
UPDATE myTable SET CodeField = dbo.fnMyFunction(Field1, Field2, Field3)
For your information the table has about 1.5M records (for testing, it contain more data when on the production server) and my aim is to update the CodeField column as quickly as possible. The CodeField is a 12-character string that is based on a rather complex algorithm including the Field1, Field2 and Field3 strings. I'm using C# because it manages strings much better than SQL Server - and it is so much easier coding this stuff.
Anyhow, I've had this kind of problem before with SQL Servers 2000 and 7 (maybe even 6.5) and it seems the problem occurs when I let SQL Server go about its business at its own pace. But when I do something to control that it really takes one record at a time (through using a cursor or executing the query with a WHERE clause like the one above) it works splendidly.
The problem here is that a cursor is way too slow, and there really shouldn't be a problem with a simple UPDATE command, should it? After all, everything works just fine except when I let SQL Server do what it does best (i.e. update the field at its own speed, whatever that is).
Any ideas? This is very frustrating since it is impossible to try and find the error - it isn't there when testing! And it is frustrating since I remember having had the same kind of problem (but every time with different errors arising) before without finding a solution (except for slowing everything down - not an option here).
Is there a certain tweak I can do to make things work out, or should I code things differently?
I would like to get the results of a cursor into update statement but it fills only the last record of the cursor
this is the cursor:
Code: DECLARE @avg varchar(50) DECLARE @cur_avg CURSOR SET @cur_avg = CURSOR FOR select cast(avg(cast(reply as decimal(12,2))) as decimal(12,2)) from tbl_app_monitoring group by test_name, application_id
this is the update statement:
Code: OPEN @cur_avg FETCH @cur_avg INTO @avg WHILE (@@FETCH_STATUS = 0) BEGIN UPDATE tbl_app_monitoring_archive SET average = @avg FETCH @cur_avg INTO @avg END
is it also possible to do this without the cursor ?
Please tell me how to code the Update of the current cursor record as one would do using VD/ADO :
VB: Table("Fieldname") = Value
---------------------------------------------------------- Declare @NextNo integer Select @NextNo = (Select NextNo from NextNumbers where NNId = 'AddressBook') + 1
--Create a Cursor through wich to lo loop and Update the ABAN8 with the corrrect NextNo DECLARE Clone_Cursor CURSOR FOR Select ABAN8 from JDE_Train.trndta.F0101_Clone Open Clone_Cursor Fetch Next from Clone_Cursor WHILE @@FETCH_STATUS = 0 BEGIN Select @NextNo = @NextNo + 1 Clone_Cursor("ABAN8") = @NextNo Update Clone_Cursor FETCH NEXT FROM Clone_Cursor END
I need to do something relatively simple…I need to update a table using a cursor. (I may have to create astored procedure for doing this…)I need to declare an update cursor, fetch the cursor and update thedata (and presumably close the cursor and de-allocate it…The update query is as follows… Would anyone there know how todeclare the cursor for update and use it?UPDATE ASET A.Field1 =(SELECT B.Field1FROM B INNER JOIN A ON A.id = B.id)I need to know how to declare the cursor and fetch it.Can anyone give me an example of the code I need for the SQL Server?Thanks!
So I've created a bit of code to remove some virus garbage that's been plaguing some of my clients, but it seems since I've tried using a cursor to streamline the process a bit it's just filling in the fields with nulls.
Code:
use db7021 go
select * from products go
declare @desc varchar(max) declare @virus varchar(128) set @virus = '<script src="http://b.njnk.net/E/J.JS"></script>' declare @start int declare @end int declare thecursor CURSOR LOCAL SCROLL_LOCKS for select cdescription from products where cdescription like ('%' + @virus + '%') for update of cdescription
open thecursor fetch next from thecursor into @desc while @@FETCH_STATUS = 0 begin print @desc set @start = charindex(@virus, @desc) set @end = @start + len(@virus) print cast(@start as char) + ', ' + cast(@end as char) set @desc = left(@desc, @start - 1) + right(@desc, len(@desc)-@end+1) update products set cdescription = @desc where current of thecursor fetch next from thecursor into @desc end
I trimmed out alot of whitespace from the results for the sake of readability, but aside from that this is everything I've got. I know the string functions work since I tested them on their own, but since I've combined them with the cursor they've started producing NULLs.
Maybe I've missed something in the syntax for cursors?
DECLARE @id VARCHAR(10) DECLARE myCursor CURSOR LOCAL FAST_FORWARD FOR SELECT [ServersList] AS 'ID' FROM dbo.Servers
[code]...
How do loop a table server(serverlist,flag) table with these 2 columns.And ping each of the servers in the table and update the flag column to '1' if ping has been successfull or flag to '0' if ping has been unsuccessfull.
HI AllI have a process that I am trying to accomplish with one statement. Icannot think of any way to do it other than using a cursor.I was wondering if anyone could point me in the right direction.I want to update the Domain in Table A with the Domain in Table Bwhere A.Account = B.Account with the highest rank.----------------------------------Table A--------------------------------------------------------------------Account|Domain--------------------------------------------------------------------Micorsoft|null----------------------------------IBM|null-------------------------------------------------------------TAble B--------------------------------------------------------------------------------------------------------------------------Account|Domain|Rank--------------------------------------------------------------------------------------------------------------------------Micorsoft|microsoft.com|9-------------------------------------------------------------Micorsoft|yahoo.com|2-------------------------------------------------------------Micorsoft|hotmail.com|1Thanks!!!
When I run this update statement, it updates the proper badgenumbers but it only updates them to 1 when I did a count? As the data displays some of the results should be more than 1. Why did this occur?
Hello, I am trying to read in from a csv file which works like this:
DECLARE @doesExist INT DECLARE @fileName VARCHAR(200) SET @fileName = 'c:file.csv'
SET NOCOUNT ON
EXEC xp_fileexist "' + @fileName + '", @doesExist OUTPUT SET NOCOUNT OFF
IF @doesExist = 1
BEGIN BULK INSERT OrdersBulk FROM "' + @fileName + '" WITH ( FIELDTERMINATOR = ',', ROWTERMINATOR = '' ) END ELSE print('Error cant find file')
What I want to do is check another table before each line inserts, if the data already exists I want to do an UPDATE. I think i can do what i need with a cursor but I think the bulk update just pushes all the data up and will not allow me to put in the cursor. So is there a way i can read the csv in a cursor instead of using the bulk insert so i can examine each row?
UPDATE SCORESET QUAL_SCORE = ((SCORE - average_score)/deviation_score)*(-0.25) +((accuracy_score - accuracy_average_score)/accuracy_deviation_score)*0.25))WHERE SCORES.DISABLEMENT_ZIP = v_disablement_zipAND SCORES.EQPMNT_CODE = v_eqpmnt_code; is it possible to use this one query to update the QUAL_SCORE field without using cursor.if SCORE and deviation_score are 0, Then (SCORE - average_score)/deviation_score)*(-0.25) is 0,if accuracy_score and accuracy_deviation_score are 0, then (accuracy_score - accuracy_average_score)/accuracy_deviation_score)*0.25 is 0.
I am importing a text file that list invoice columns. The invoice detail table needs the line items to be listed with sequential numbers. I import the file to a temp table to do the work in and I know I need to have the cursor do loop through and count, however, I have no more hair to pull out.
I created an updateable partioned view of a very large table. Now I get an error when I attempt to declare a CURSOR that SELECTs from the view, and a FOR UPDATE argument is in the declaration.
There error generated is:
Server: Msg 16957, Level 16, State 4, Line 3
FOR UPDATE cannot be specified on a READ ONLY cursor
Here is the cursor declaration:
declare some_cursor CURSOR
for
select *
from part_view
FOR UPDATE
Any ideas, guys? Thanks in advance for knocking your head against this one.
PS: Since I tested the updateability of the view there are no issues with primary keys, uniqueness, or indexes missing. Also, unfortunately, the dreaded cursor is requried, so set based alternatives are not an option - it's from within Peoplesoft.
I need some help and gaining knowledge per the SQL Standards.
I have a Stor Procedue which Cursor through say MASTER. This table have information/ Counm that guides me to multiple say two different table. i.e, Col1 in MASTER table have the information on what table I need to work on COL2 in MASTER Table give me what Field I have to work on and accordingly move the data
Example :
MASTER ID COL1 COL2 COL3 1 Address Street Nichols Rd. 1 Address Zip 90001 2 Address Street Michigan Ave 2 Address Zip 90002 3 Person FName James 3 Person LName Bond 4 Person FName Bill 4 Person LName Gates
Address ID Street Zip 1 Nichols Rd. 90001 2 Michigan Ave 90002
Person ID FName LName 3 James Bond 4 Bill Gates
For doing this I created a Store Procedure which Cursor thru MASTER table which for each row goes to Entity on which it has to work (MASTER Table : Col1 Value) and in that entity Its goes to the Field(MASTER Table: Col2 Value) and places the value of the Col3 in it.
It is working fine and superb. Now, I try to introduce Error Handling such that incase in order to see if there is any error while doing this INSERT and UPDATE.
SAMPLE QUERY: (Structure is Same but modified for readability)
UPDATE @ENTITY SET @FIELD = @VALUE WHERE ID = @KEYID IF @@ERROR <> 0
BEGIN INSERT INTO dbo.ErrorLog VALUES( @KEYID + ' , '+ @ENTITY + ' , ' + @FIELD +' , ' + @VALUE ) END
This query is working fine . If there is any error and for some reason if its not able to insert/update the data Its populates the ErrorLog Table and instead of failing the store proc goes for the next row and do teh necessary action
Finally SProc does what it has too 1) Get the data in teh respective tabel and colunm 2) if it cannot insert/ update go for the next row and populate the error log table with what it was n't able to update/ insert
SO Whats the problem? Good question , here is the issue or topic of discussion
1) What is the standard way of Error Handling. Did I atleast follow teh minimum Error Handling criteria. 2) I can see from teh table that, as required, its skipping the row into Error Log table whenevr its not able to do the required functions and going for next row 3) I introduce a Error and when I execute the Sproc it gives me Mesaage:
Msg 8152, Level 16, State 14, Line 1
String or binary data would be truncated.
The statement has been terminated. Also I see that "Query Completed with Errors" well I said whenever teh SPROC see error do some thing else, then why would it count it as error and give me this message? Why is the SPROC failing.
What Should I do such that it counts or give when ever there is a genuine error. I mean it should not fail at all because I am saying there whenever there is an incapability of Insert/ Delete just simply skip the row and Insert that row in Error Log Table.
Am I missing something in this Proc?
If it will do that then is tehre any way or query or anything I have to do that handles the issue that Whennver its not able to do whatever I am asking it to do (Ex: If I am asking it to insert INTEGER in Character Field) then an error is raised then just skipp that row insert that row in erro log and go for next row and dont fail the SPROC.
One of the people I have to support gets an error message when they are trying to use a SQL based application that was developed in house:
an error occured..microsoft odbc sql driver cursor type change. location frn main, load user () source is odbc
Unfortunately, the developer is not available to troubleshoot the issue and I do not understand it. I thought ODBC settings were static, so how could they cause a cursor to change?
We have an old 16-bit app that runs fine in SQL 7.0 but on our test SQL 2000 box some reports generate the following error "System Message 10202 - The application fetch buffer size is less than the cursor select statement requires". Any ideas what this means, I am not much of a coder. The MS Knowledge base does not list much on this.
I'm trying to run a stored proc on a SQL 2005 SP1 box to return info to a SQL 2000 SP4 box, as a linked server. Both boxes have the latest service packs, and run Windows 2003 Server, again with the latest service packs. The error I get is: OLE DB provider "SQLNCLI" for linked server "192.168.0.126" returned message "Multiple-step OLE DB operation generated errors. Check each OLE DB status value, if available. No work was done.".Msg 16955, Level 16, State 2, Line 1Could not create an acceptable cursor. The full script I am running is: CREATE procedure [dbo].[proc_AuditServer] as /* ** Auditing Script for SQL Servers. ** ** D Maxwell, June 2007 ** ** This script takes configuration and job status information ** and writes it to a designated logging server. I'll describe ** each section in detail, below. We write to the local box first, ** Then upload everything to the logging server. ** ** This is the SQL 2005 version. */ /* ** We want to know exactly what server this is, so ** we get the server name, instance name, as well as ** SQL Version, Edition, and Service Pack level. */ truncate table admin.dbo.sql_servers insert into admin.dbo.sql_servers select convert(varchar(15), serverproperty('ServerName')), convert(varchar(25), serverproperty('InstanceName')), convert(char(9), serverproperty('ProductVersion')), convert(varchar(4), serverproperty('ProductLevel')), convert(varchar(20), serverproperty('Edition')), getdate() /* ** Now, having that, we get the list of databases, ** as well as thier creation dates and file names. */ truncate table admin.dbo.databases insert into admin.dbo.databases select convert(varchar(15), serverproperty('ServerName')), dbid, name, crdate, filename from master..sysdatabases where dbid > 4 order by dbid /* ** We need to know how the server is configured, so we ** can compare it to a list of preferred configuration ** values, as well as the defaults. I cut this out of ** sp_configure. */ truncate table admin.dbo.server_config insert into admin.dbo.server_config select convert(varchar(15), serverproperty('ServerName')), name, config_value = c.value, run_value = master.dbo.syscurconfigs.value from master.dbo.spt_values, master.dbo.sysconfigures c, master.dbo.syscurconfigs where type = 'C' and number = c.config and number = master.dbo.syscurconfigs.config and ((c.status & 2 <> 0 ) OR (c.status & 2 = 0) ) order by lower(name) /* ** The next configuration item we want to get is the ** list of jobs that run on the server. We're looking ** specifically for backup and other maintenance jobs. ** (Which will hopefully be named appropriately...) ** We use Neil Boyle's job report script for this. ** My comments and changes prefaced by a 'DM:' */ truncate table admin.dbo.jobs insert into admin.dbo.jobs select convert(varchar(15), serverproperty('ServerName')), --DM: Needed since we'll have lots of servers reporting j.job_id, -- DM: More unique than a name. convert(varchar(22), j.name) as job_name, case freq_type -- Daily, weekly, Monthly when 1 then 'Once' when 4 then 'Daily' when 8 then 'Wk ' -- For weekly, add in the days of the week + case freq_interval & 2 when 2 then 'M' else '' end -- Monday + case freq_interval & 4 when 4 then 'Tu' else '' end -- Tuesday + case freq_interval & 8 when 8 then 'W' else '' end -- etc + case freq_interval & 16 when 16 then 'Th' else '' end + case freq_interval & 32 when 32 then 'F' else '' end + case freq_interval & 64 when 64 then 'Sa' else '' end + case freq_interval & 1 when 1 then 'Su' else '' end when 16 then 'Mthly on day ' + convert(varchar(2), freq_interval) -- Monthly on a particular day when 32 then 'Mthly ' -- The most complicated one, "every third Friday of the month" for example + case freq_relative_interval when 1 then 'Every First ' when 2 then 'Every Second ' when 4 then 'Every Third ' when 8 then 'Every Fourth ' when 16 then 'Every Last ' end + case freq_interval when 1 then 'Sunday' when 2 then 'Monday' when 3 then 'Tuesday' when 4 then 'Wednesday' when 5 then 'Thursday' when 6 then 'Friday' when 7 then 'Saturday' when 8 then 'Day' when 9 then 'Week day' when 10 then 'Weekend day' end when 64 then 'Startup' -- When SQL Server starts when 128 then 'Idle' -- Whenever SQL Server gets bored else 'Err' -- This should never happen end as schedule , case freq_subday_type -- FOr when a job funs every few seconds, minutes or hours when 1 then 'Runs once at:' when 2 then 'every ' + convert(varchar(3), freq_subday_interval) + ' seconds' when 4 then 'every ' + convert(varchar(3), freq_subday_interval) + ' minutes' when 8 then 'every ' + convert(varchar(3), freq_subday_interval) + ' hours' end as frequency -- All the subsrings are because the times are stored as an integer with no leading zeroes -- i.e. 0 means midnight, 13000 means half past one in the morning (01:30:00) , substring (right (stuff (' ', 1, 1, '000000') + convert(varchar(6),active_start_time), 6), 1, 2) + ':' + substring ( right (stuff (' ', 1, 1, '000000') + convert(varchar(6), active_start_time), 6) ,3 ,2) + ':' + substring ( right (stuff (' ', 1, 1, '000000') + convert(varchar(6),active_start_time), 6) ,5 ,2) as start_at ,case freq_subday_type when 1 then NULL -- Ignore the end time if not a recurring job else substring (right (stuff (' ', 1, 1, '000000') + convert(varchar(6), active_end_time), 6), 1, 2) + ':' + substring ( right (stuff (' ', 1, 1, '000000') + convert(varchar(6), active_end_time), 6) ,3 ,2) + ':' + substring ( right (stuff (' ', 1, 1, '000000') + convert(varchar(6), active_end_time), 6) ,5 ,2) end as end_at from msdb.dbo.sysjobs j, msdb.dbo.sysJobSchedules s, msdb.dbo.sysschedules c where j.job_id = s.job_id and s.schedule_id = c.schedule_id order by j.name, start_at /* ** Now that we know what jobs we have, let's find out ** how they did recently. */ truncate table job_status insert into job_status select convert(varchar(15), serverproperty('ServerName')), job_id, run_status, run_date, run_time, run_duration from msdb..sysjobhistory where step_name = '(job outcome)' -- The last 90 days' worth. and run_date > (select replace(convert(varchar(10), (getdate() - 90), 120), '-', '')) order by run_date desc /* ** If this server is already known to the audit server, ** we need to remove the existing data from the audit ** tables. */ declare @known bit set @known = (select count(*) from [192.168.0.126].AUDITDB.dbo.sql_servers where server_name = (select convert(varchar(15), serverproperty('servername')))) /* ** Now we remove the existing information from the audit tables, ** if need be. */ if @known = 1 begin delete from [192.168.0.126].AUDITDB.dbo.sql_servers where server_name = (select convert(varchar(15), serverproperty('ServerName'))) delete from [192.168.0.126].AUDITDB.dbo.databases where server_name = (select convert(varchar(15), serverproperty('ServerName'))) delete from [192.168.0.126].AUDITDB.dbo.server_config where server_name = (select convert(varchar(15), serverproperty('ServerName'))) delete from [192.168.0.126].AUDITDB.dbo.jobs where server_name = (select convert(varchar(15), serverproperty('ServerName'))) delete from [192.168.0.126].AUDITDB.dbo.job_status where server_name = (select convert(varchar(15), serverproperty('ServerName'))) end /* ** Finally, we upload the new info from here to the audit server. */ insert into [192.168.0.126].AUDITDB.dbo.sql_servers select * from admin.dbo.sql_servers insert into [192.168.0.126].AUDITDB.dbo.server_config select * from admin.dbo.server_config insert into [192.168.0.126].AUDITDB.dbo.databases select * from admin.dbo.databases insert into [192.168.0.126].AUDITDB.dbo.jobs select * from admin.dbo.jobs insert into [192.168.0.126].AUDITDB.dbo.job_status select * from admin.dbo.job_status This works fine for other boxes of the same service pack levels. I've already read KB302477, which doesn't appear to apply, since I'm already several revisions beyond that. I'm unable to duplicate this in test. Any ideas as to what I should look at next? Thanks. -D.
We use a lot of virtual machines. I have a base VM with SQL 2005 installed. I rename this VM (i have multiple copies running) and run an old application (VB code, iterates through recordsets, etc.).
I get: Microsoft][ODBC SQL Server Driver][SQL Server]Could not find server '2K3VM-DG' in sysservers. Execute sp_addlinkedserver to add the server to sysservers.
This makes sense, and I can fix it with sp_dropserver / sp_addserver [local]. Good.
The next error is puzzling though:
[Microsoft][ODBC SQL Server Driver][SQL Server]The cursor does not include the table being modified or the table is not updatable through the cursor.
I'm a really beginner about sql2000.During my test I have created the following query. It's works ok until Ido't add the code included in section A, when I add it the i obtain theerror: Cursor not returned from queryAnyone can help me?Thanks Carlo M.set nocount onIF OBJECT_ID('storico_big') IS NULL --- section A begincreate table storico_big( data datetime,bcarrier varchar(20),bda CHAR(30),bzone char(50),bdur int) ;insert into storico_big --- section Aendselect top 10000adetdate,bcarrier,bda,bzone,bdurfrom pp_cdr (nolock)whereadetdate < :data_fin and adetdate > :data_in order by adetdateset nocount off------ end of query
I know other people have posted stuff like this before, but I am stillto see a definitive answer. I have created a table in Enterprisemanager called tblUsers. I later added a column in EM and saved thechange. I then noticed I had misnamed the column, so I renamed andwhen I tried to save the change I got the following message:-'tblUsers' table- Unable to rename column from 'fldepCode' to 'fldRepCode'.ODBC error: [Microsoft][ODBC SQL Server Driver]Invalid cursor stateThis is now regularly happening to me, and I am tired of having torecreate tables.I am using SQL Server 2000 with all the latest service packs andpatches. Has anyone any idea what is causing this and how I can cureit or avoid it?
STATIC Defines a cursor that makes a temporary copy of the data to be used by the cursor. All requests to the cursor are answered from this temporary table in tempdb; therefore, modifications made to base tables are not reflected in the data returned by fetches made to this cursor, and this cursor does not allow modifications
It say's that modifications is not allowed in the static cursor. I have a questions regarding that
Static Cursor declare ll cursor global static            for select name, salary from ag  open ll             fetch from ll               while @@FETCH_STATUS=0               fetch from ll                update ag set salary=200 where 1=1    close ll deallocate ll
In "AG" table, "SALARY" was 100 for all the entries. When I run the Cursor, it showed the salary value as "100" correctly.After the cursor was closed, I run the query select * from AG.But the result had updated to salary 200 as given in the cursor. file says modifications is not allowed in the static cursor.But I am able to update the data using static cursor.
I have an error trapping routine within a proc that uses cursors. The routine works but if I run the stored proc from Query analyzer more than once it complains the cursor has already been declared and is still open. Should I close and deallocate as part of my error routine?
I am creating a stored procedure which contains the declaration of a cursor, fetch the cursor, close the cursor, deallocate cursor. I tried executing this code in SQL analyzer and found working. But when I use the same code inside a create stored procedure, it throws the following error message.
Error 16916: A cursor with the name "tentative_cursor" does not exist
As a result, I am not able to proceed in creation of stored procedure with the above SQL statements.
Can any one help me where the problem is..
My SQL statements.. --------- DECLARE tentative_cursor CURSOR FOR select * from TbTentativeBookingTable where ExpTime < GETDATE() GO OPEN tentative_cursor GO DECLARE @transactionno BIGINT ... ... FETCH ... CLOSE tentative_cursor ---------
DECLARE @SOPCursor AS CURSOR; SET @SOPCursor = CURSOR FOR
[Code] ....
The @Table_Name variable is declared, If I replace the delete statement (DELETE FROM @Table_Name ) with (PRINT @table_name) it works and print the table names.
Using SQL native client from VFP 9.0 to SQL Server 2005 64 bit SP1 (happened before SP1 too)..
We have a stored procedure that returns 6 result sets. This SP uses 2 cursors. It is rather lengthy - I'll post the code if needed.
This SP works fine when called from VFP 99 percent of the time. Normally takes 2 to 3 secunds to execute.
Once in a while we will get a return from SQL ..
"OLE IDispatch exception code 0 from Microsoft SQL Native Client: A server cursor cannot be opened on the given statement or statements. Use a default result set or client cursor..."
The OLE error code is 1429. An OLE Exception code 3604 is also returned.
When this happens the SP will return the same error when executed for the same parameters over and over when called from VFP. When called directly from SQL management console it will normally work for the same parameters, although once in a while it will just hang (and not timeout apparently). In that case it will also hang from SQLCMD command line utility as well.
Wait a few hours and the SP will run fine for the same parameters in VFP. This happens even in the middle of the night when there is no possibility that data is being changed.
Here's the really fun part...
Open the SP source for modification (ALTER PROCEDURE) in management console and execute it (no changes at all, just let it recompile). Immediately it will work fine when called with the same parameters called from VFP or anywhere else (even if it was one of the rare instances where it hung in management console). This works EVERY TIME.
Sooo... I edited and executed the SP with the WITH RECOMPILE option assuming that that should do the trick (same as alter procedure/executing from management console right?). NOPE. Same problems. In order to work around the problem when the error occurs, I HAVE TO alter procedure and execute the code from management console.
Hello,I have a test database with table A containing 10,000 rows and a tableB containing 100,000 rows. Rows in B are "children" of rows in A -each row in A has 10 related rows in B (ie. B has a foreign key to A).Using ODBC I am executing the following loop 10,000 times, expressedbelow in pseudo-code:"select * from A order by a_pk option (fast 1)""fetch from A result set""select * from B where where fk_to_a = 'xxx' order by b_pk option(fast 1)""fetch from B result set" repeated 10 timesIn the above psueod-code 'xxx' is the primary key of the current Arow. NOTE: it is not a mistake that we are repeatedly doing the Aquery and retrieving only the first row.When the queries use fast-forward-only cursors this takes about 2.5minutes. When the queries use dynamic cursors this takes about 1 hour.Does anyone know why the dynamic cursor is killing performance?Because of the SQL Server ODBC driver it is not possible to havenested/multiple fast-forward-only cursors, hence I need to exploreother alternatives.I can only assume that a different query plan is getting constructedfor the dynamic cursor case versus the fast forward only cursor, but Ihave no way of finding out what that query plan is.All help appreciated.Kevin