How To Capture SQL Server Does Not Exist Type Errors Within Stored Procs
Nov 16, 2007
Hi, I have a stored procedure running on our monitoring server to monitor the backup jobs on other sql servers running on our network. This sp loops thorugh the list of server names and connects remotely to other servers using linked servers. However, if one of the servers is down, the sp stops after raising 42000 error ("SQL Server does not exist") and fails to carry on processing the next server in the list.
Is there any way to capture this "SQL Server does not exist" error within the sp so that the sp can carry on with the processing?
I am using SQL Server 2000 SP4.
Thanks in advance for any replies.
Opal
Ok, I've read somewhere(which I'm looking for again : ) that said that there are errors like DeadLock that kills the execution of a stored proc and there are other errors that do not necessarily kill the rest of the execution of the stored proc. Is that true? If so does anyone have any links I can read. What I'm seeing is a bad id in the foreign key and I think what is happening is that there was a unique constraint error on the first insert but the stored proc continued executing and used the bad id later on in the stored proc.
I do know I can use the @@error and will start using it but I need more proof to agree or not agree with my theory.
Thanks ahead of time for any information you can give me either way.
public Set DoSomthing(Set toBeProcessed, Set measuresToWorkWith)The set measurseToWorkWith is passed as {[Measures].[Measure1], [Measures].[Measure2] ...}
with the measures being real or query-scoped calculated members.
To get the value of the measure for each tuple in the set toBeProcessed, I create an Expression for each tuple (measure) in the set measuresToWorkWith then for each tuple in toBeProcessed call expression.Calculate(tuple) which returns a MDXValue.
My problem is that in order to make the code generic I need to get the real (.NET) data type of the MDXValue. The class only has explicit conversion methods ToInt16() etc which implies that the data type is known at design time.
However, if one of the measures is a query-scoped calculation then it could return a .NET double, int, bool or string.
If the measure is real then I can look up its metadata. However, it appears that if it is a formula (scoped member) then are all bets are off?
I have a stored procedure that dynamically bulk loads several tables from several text files. If I encounter an error bulk loading a table in the stored procedure, all I get is the last error code produced, but if I run the actual bulk load commands through SQL Management Studio, it gives much more usable errors, which can include the column that failed to load. We have tables that exceed 150 columns (don't ask), and having this information cuts troubleshooting load errors from hours down to minutes. Onto my question..., is there any way to capture all of the errors produced by the bulk load from within a stored procedure (see examples below)?
Running this...
BULK INSERT Customers
FROM 'c: estcustomers.txt'
WITH (TabLock, MaxErrors = 0, ErrorFile = 'c: estcustomers.txt.err')
Produces this (notice column name at the end of the first error)...
Msg 4864, Level 16, State 1, Line 1
Bulk load data conversion error (type mismatch or invalid character for the specified codepage) for row 1, column 1 (CustId).
Msg 7399, Level 16, State 1, Line 1
The OLE DB provider "BULK" for linked server "(null)" reported an error. The provider did not give any information about the error.
Msg 7330, Level 16, State 2, Line 1
Cannot fetch a row from OLE DB provider "BULK" for linked server "(null)".
Running this (similar to code in my stored procedure)...
BEGIN TRY
BULK INSERT Customers
FROM 'c: estcustomers.txt'
WITH (TabLock, MaxErrors = 0, ErrorFile = 'c: estcustomers.txt.err')
END TRY
BEGIN CATCH
SELECT
ERROR_NUMBER() AS ErrorNumber,
ERROR_SEVERITY() AS ErrorSeverity,
ERROR_STATE() AS ErrorState,
ERROR_PROCEDURE() AS ErrorProcedure,
ERROR_LINE() AS ErrorLine,
ERROR_MESSAGE() AS ErrorMessage;
END CATCH
Produces something similar to this (which is useless)... ...Cannot fetch a row from OLE DB provider "BULK" for linked server "(null)".
When a stored PL/SQL procedure in my Oracle database is called from ASP.NET, is it possible to retrieve the OUT parameter from the PL/SQL procedure? For example, if I have a simple procedure as below to insert a row into the database. Ideally I would like it to return back the parameter named NewId to my ASP.NET server. I'd like to capture this in the VB.NET code. 1 create or replace procedure WriteName(FirstName in varchar2, LastName in varchar2, NewId out pls_integer) is2 3 NameId pls_integer;4 5 begin6 7 select name_seq.nextval into NameId from dual;8 9 insert into all_names(id, first_name, last_name)10 values(NameId, FirstName, LastName);11 12 NewId := NameId;13 14 end WriteName; 1 <asp:SqlDataSource 2 ID="SqlDataSaveName" 3 runat="server" 4 ConnectionString="<%$ ConnectionStrings:ConnectionString %>"5 ProviderName="<%$ ConnectionStrings:ConnectionString.ProviderName %>" 6 SelectCommand="WRITENAME"7 SelectCommandType="StoredProcedure">8 <SelectParameters>9 <asp:ControlParameter ControlID="TextBoxFirstName" Name="FIRSTNAME" PropertyName="Text" Type="String" />10 <asp:ControlParameter ControlID="TextBoxLastName" Name="LASTNAME" PropertyName="text" Type="String" />11 </SelectParameters>12 </asp:SqlDataSource>This is then called in the VB.NET code as below. It is in this section that I would like to capture the PL/SQL OUT parameter NewId returned from Oracle. 1 SqlDataSaveName.Select(DataSourceSelectArguments.Empty) If anybody can help me with the code I need to add to the VB.NET section to capture and then use the returned OUT parameter then I'd be very grateful.
Is there a way to namespace groups of stored procs to reduce confusion when using them?
For C# you can have ProjectName.ProjectSection.Classname when naming a class. I am just wondering if there is a way to do the same with SQL stored procs. I know Oracle has packages and the name of the package provides a namespace for all of the stored procs inside it.
I currently have a stored procedure that is defined as follows:
CREATE PROCEDURE UpdateSyncLog
@TableName char(100),
@LastSyncDateTime datetime,
@ErrorState int OUTPUT
I am using an execute sql task to call this procedure. The connectiontype is ADO .NET and the SQLSourceType is DirectInput. The IsQueryStoredProcedure setting is false, and the following is my SQL Statement I have entered:
Result set is set to None, as this query returns NO results (i.e. has no select statements in it that returns results).
I have two variables in this SSIS package. CurrentDateTime, and ErrorStateVal. CurrentDateTime is of Data type DateTime, the ErrorStateVal is of type Int32
The error I am getting when running this execute sql task is as follows:
Error: 0xC001F009 at AS400 to SQL Full Repopulation Sync: The type of the value being assigned to variable "User::ErrorStateVal" differs from the current variable type. Variables may not change type during execution. Variable types are strict, except for variables of type Object.
Error: 0xC002F210 at Execute SQL Task, Execute SQL Task: Executing the query "exec UpdateSyncLog 'myTestTable', @LastSyncDateTime, @ErrorState" failed with the following error: "The type of the value being assigned to variable "User::ErrorStateVal" differs from the current variable type. Variables may not change type during execution. Variable types are strict, except for variables of type Object.
". Possible failure reasons: Problems with the query, "ResultSet" property not set correctly, parameters not set correctly, or connection not established correctly.
Task failed: Execute SQL Task
This makes no sense to me, both the SSIS variable ErrorStateVal is Int32, as well as the parameter declaration in the Execute SQL task is Int32 with direction of OUTPUT, and my stored procedure definition has @ErrorState as an integer as well.
I use a CLR external function to call a payment gateway, for some reason, I could not get any status sometime, check the following code, I expect I get either -2 or -3 in status no matter what error happens - include the gateway error or the CLR error since I have try catch there. Do you think the following code could miss any problem?
update link_payments set ccagentcode=@ccagentcode,status=-2,postresult=@postresult where id=@paymentid
.... end try begin catch select @message= '<br>ErrorNumber: '+cast(ERROR_NUMBER() as varchar)+char(13)+ '<br>ErrorSeverity: '+cast(ERROR_SEVERITY() as varchar)+char(13)+ '<br>ErrorState: '+cast(ERROR_STATE() as varchar)+char(13)+ '<br>ErrorProcedure: '+isnull(ERROR_PROCEDURE(),'')+char(13)+ '<br>ErrorLine: '+cast(ERROR_LINE() as varchar)+char(13)+ '<br>ErrorMessage: '+isnull(ERROR_MESSAGE(),'')+char(13)
update link_payments set status=-3,postresult=@message where id=@paymentid return -1 end catch
I have created a package which executes every 10 mins. Last week end for maintenance purpose, I shutdown my database. Now as an initial execution process, my package does the default validation steps on which the database connection validation step fails. As this is the default functionality of SSIS I am not able to capture this error. Is there anyway to capture this error inside SSIS Package?
I have an Execute Process Task within my package that executes a BCP command which outputs the resultset of a query to a file in the network share. It works fine most of the time, however sometimes for unknown reasons, the following error message gets logged in my log table - In Executing "c:Program FilesMicrosoft SQL Server90ToolsBinncp.exe" "Select Comments, SoldToCustomerNbr, ProductGroupingCode, ProductGroupingName, RevSumCategoryCode, RevSumCategoryName, ValidFromDate, DTSCollectPct, DTSPrepaidPct, DTSPickUpPct, DCCollectPct, DCPrepaidPct, DCPickUpPct From ShipmentTypeCustomerBlend" queryout \xxxLOGShipmentTypeCustomerBlendLog_060719201440.txt -c -t" " -SDummyServer -T -e"d:SSIS Error LogsJob ExecutionBcpErrors.log" at "", The process exit code was "1" while the expected was "0".
The above error was captured from the System::ErrorDescription variable, by the error event handler, that was attached to the Execute Process Task. This error does not help me to debug the issue.
On running the below statement from the command prompt, i get the actual error message, which is the expected behavior -
This error message indicates that either the network path - \xxxLOG is not available for the output file creation or the file - \xxxLOGShipmentTypeCustomerBlendLog_060719201440.txt could not be created for some reason.
I 've tried to capture the error message from the StandardErrorVariable and the StandardOutputVariable properties of the ExecuteProcess Task, but in vain.
Is this a bug ? If so, is there a way to get the actual error message from the task ?
Is it possible to capture the errors which raising in System,Databases,networking using SQLServer UDB(assume DBA Database). or Is it possible to store all Errors(system,Databases,networking...Services failure...)in SQL Server UDB.
Recently I have experienced the following scary scenario;
Users report a problem with an app, on diagnosis it is found that a bunch of SPs in SQL Server seem to be corrupt.
That is, when the code for them is listed, the 'Create Procedure..' line and several DECLARE lines are missing. Often the SP starts off in the middle of a variable name!!
Perhaps entries in the syscomments table have become corrupted?
Has anyone seen this problem before - is it perhaps a known bug in SQL Server 6.5?
CREATE PROCEDURE PROC1 AS BEGIN SELECT A.INTCUSTOMERID,A.CHREMAIL,B.INTPREFERENCEID,C.CHR PREFERENCEDESC FROM CUSTOMER A INNER JOIN CUSTOMERPREFERENCE B ON A.INTCUSTOMERID = B.INTCUSTOMERID INNER JOIN TMPREFERENCE C ON B.INTPREFERENCEID = C.INTPREFERENCEID WHERE B.INTPREFERENCEID IN (6,7,2,3,12,10) ORDER BY B.INTCUSTOMERID
END
IF I AM USING THIS PROC AS I/P TO ANOTHER PROC THEN IT GIVES NO PROBLEM AS I CAN USE ?
BUT IF , I USE ANOTHER PROC SIMILAR TO THE FIRST ONE WHICH GIVES SLIGHTLY DIFFERENT RESULTS AND GIVES TWO SETS OF RESULTS,THEN WE HAVE A PROBLEM,HO TO SOLVE THIS :-
CREATE PROCEDURE MY_PROC AS BEGIN SELECT A.INTCUSTOMERID,A.CHREMAIL,B.INTPREFERENCEID,C.CHR PREFERENCEDESC FROM CUSTOMER A INNER JOIN CUSTOMERPREFERENCE B ON A.INTCUSTOMERID = B.INTCUSTOMERID INNER JOIN TMPREFERENCE C ON B.INTPREFERENCEID = C.INTPREFERENCEID WHERE B.INTPREFERENCEID IN (23,12,10) ORDER BY B.INTCUSTOMERID
END
SELECT A.INTCUSTOMERID,MAX(case when A.intpreferenceid = 23 then '1' else '0' end) + MAX(case when A.intpreferenceid = 12 then '1' else '0' end) + MAX(case when A.intpreferenceid = 10 then '1' else '0' end) AS PREFER FROM CUSTOMER GROUP BY A.INTCUSTOMERID ORDER BY A.INTCUSTOMERID END
WHICH NOW GIVES ME TWO SETS OF DATA , BOTH REQUIRED THEN HOW TO USE ONE SET OF RESULTS AS I/P TO ANOTHER PROC AND OTHER SET OF RESULTS AS I/P TO YET ANOTHER PROC .
I am trying to create a SQL data adapter via the wizard, however, I get the error "Invalid object name #ords" because the stored procedure uses a temp table. Anyway around this? Thanks.
I would like to capture any errors thrown by linked server in my stored procedure, How do I do this? eg: Select statement in sp to remote server fails if login credentials were changed for some reason on remote server. @@error is not able to capture this error and stored procedure is terminated when this type of error occurs.
We have a product that we were experiencing high recompile rates do totemp tables, we have re-written them to use table variables to reducerecompile rate. The problem I am having is that we have somecustomers still on SQL 7 where the table variables were not valid. Iwould like to have one installation script that would create thestored procedure with table variables on SQL Server 2000 and withouton SQL Server 7, I have been unsuccessful do to the limitation thatCREATE PROCEDURE statement cannot be in a script with any otherstatements.Was wondering if anyone else has run into this and found a way aroundit.
I have around 100 packages (all [packages run at same time) each package calls a stored Procedure once the Stored proc Execution is completed the package will write the log information into a log table, here how can i capture the CPU and Memory usage for execution of each stored proc.
If you drop / rename a table and then recreate the table with the SAME NAME, what impact does it have on stored procedures that use these tables? From a system perspective, do you have to rebuild / recompile ALL the stored procedures that use this table?
I had a discussion with someone that said that this is a good idea, since the IDs of the tables change in sysobjects and from a SQL SERVER query plan perspective, this needs to be done...
Question 2
If you Truncate a Table as part of a BEGIN TRANSACTION, what happens if an error occurs? Will it Rollback? The theory is that it won't because Truncate doesn't utilize the logs where as Delete From uses the SQL Logs?
I work with a large and complex reporting system with several hundred reports: the Programmability; Stored Procedures node of the object explorer has become very difficult to navigate.
Is there any metadata that can be embedded in the stored procs that would create subfolders like the existing System Stored Procedures node in this node of the object explorer?
I suspect that the correct answers are:
Rename all your queries with a rational naming convention; Cull the deadwood; Assign them to categories then export them to separate databases.
Unfortunately, one of these has already been done, and the other two will break several hundred dependent processes - the recoding and retesting is neither economical nor desirable.
Still, all advice is welcome. Pitch your answers at a banking geek who does intermediate to advanced stored procedures and triggers, but isn't allowed to play with sharp things (like sys objects) - but I can probably get help from a grown-up on the sysadmin team: I have discovered that the rumours about human sacrifice are baseless, and they will perform favours in return for beer.
This is also a good time to ask: just how many stored procs and functions are you allowed in SQL Server 2005?
How do I search for and print all stored procedure names in a particular database? I can use the following query to search and print out all table names in a database. I just need to figure out how to modify the code below to search for stored procedure names. Can anyone help me out? SELECT TABLE_SCHEMA + '.' + TABLE_NAME FROM INFORMATION_SCHEMA.TABLES WHERE TABLE_TYPE = 'BASE TABLE'
I've got a big problem that I'm trying to figure out:I have an address table out-of-which I am trying to select mailing addresses for companies UNLESS a mailing address doesn't exist; then I want to select the physical addresses for that company. If I get multiple mailing or physical addresses returned I only want the most recently edited out of those.I don't need this for an individual ID select, I need it applied to every record from the table.My address table has some columns that look like: [AddressID] [int][LocationID] [int][Type] [nvarchar](10)[Address] [varchar](50)[City] [varchar](50)[State] [char](2)[Zip] [varchar](5)[AddDate] [datetime][EditDate] [datetime]AddressID is a primary-key non-null column to the address table and the LocationID is a foreign key value from a seperate Companies table.So there will be multiple addresses to one LocationID, but each address will have it's own AddressID.How can I do this efficiently with perfomance in mind???Thank you in advance for any and all replies...
I've got a big problem that I'm trying to figure out:
I have an address table out-of-which I am trying to select mailing addresses for companies UNLESS a mailing address doesn't exist; then I want to select the physical addresses for that company. If I get multiple mailing or physical addresses returned I only want the most recently edited out of those.
I don't need this for an individual ID select, I need it applied to every record from the table.
My address table has some columns that look like: [AddressID] [int] [LocationID] [int] [Type] [nvarchar](10) [Address] [varchar](50) [City] [varchar](50) [State] [char](2) [Zip] [varchar](5) [AddDate] [datetime] [EditDate] [datetime]
AddressID is a primary-key non-null column to the address table and the LocationID is a foreign key value from a seperate Companies table. So there will be multiple addresses to one LocationID, but each address will have it's own AddressID.
How can I do this efficiently with perfomance in mind???
I am wondering if it is possible to execute an .exe from a stored proc. I have an application that calls a stored proc and I am wanting for that stored proc to call up an app to show certain results. I have no way of doing this on the application itself as I dont have the source code. So, since I do have access to the SQL I am wondering if it could be done there.
First off, this is a cross post which is also located here: http://www.sqlteam.com/forums/topic.asp?TOPIC_ID=34073
Now, with the following stored procs and UDF, the Result is NULL and I cannot figure out why.
What I am doing is calling a stored procedure pass in an OUTPUT variable of type float. This stored proceudre then calls a different stored procedure passing the same OUTPUT variable of type float. The second stored procedure then calls a UDF passing in two variables to be multiplied ans this should set the OUTPUT variable to the result.
The UDF does return the correct result to the interior stored procedure, but the interior stored procedure does not pass the value back to the original stored procudure resultint in a Result value = NULL.
The code is below, just copy / past, execute and you will see exactly what I mean.
Any thoughts?
USE NORTHWIND GO
CREATE FUNCTION RECTANGULAR_XSECTION (@fWidth float, @fHeight float)
I am trying to improve my SQL and reduce the number of connections my website is making to the server.
Currently I have a stored procedure that gets all the games (and their details) for a single user which works fine which uses a WHERE userid = @userid which is an Int value.
I now need to create a new procedure which brings back the to all their "friends" (again by their userid). How is it best to do this? I originally tried to do a WHERE userID IN (@userid) but was unable to work out what variable type to use for @userid.
Is it best to do a single query in this way or is there a way to use the existing SP, loop through the collection of userids and join the result together into a single set to return? If an IN is the best route, what is the correct datatype for the variable?
can anyone help me out.... i need to compare a date that was retrieved from the database and the system date then it must be coded in the stored procs of the database.. help!!!!
I'm relatively new to stored procs (not to SQL or SQL Server) and I am trying to get transactions to work within a stored proc. Here is the code:
( --define the parameters that are needed @userID char(32), @userName varchar(50), @status char(10), @type char(10), @password varchar(50), @firstName varchar(100), @lastName varchar(100), @email varchar(200), @domain varchar(50), @pwdExpiry int, @badLogin int, @lastLogin datetime, @full varchar(8000), @read varchar(8000), @noaccess varchar(8000) ) AS BEGIN TRAN tmp1 --First, insert the user into the system INSERT INTO ptsUsers(UserID, UserName, Status, Type, Password, FirstName, LastName, Email, DomainName, PasswordExpiry, BadLoginAttempts, LastLoginDate) values(@userID, @userName, @status, @type, @password, @firstName, @lastName, @email, @domain, @pwdExpiry, @badLogin, @lastLogin);
--Now, we need to add the function access edges for the user declare @arrayValue char(32) declare @rightID char(32) declare @sepPos int while patindex('%,%',@full)<>0 BEGIN select @sepPos = patindex('%,%' , @full) select @arrayValue = left(@full, @sepPos - 1) -- replace the value with an empty string select @full = stuff(@full, 1, @sepPos, '')
--create and parse the new id declare @strID char(32),@tmpStr varchar(40) set @tmpStr = newid(); set @strID = REPLACE(@tmpStr,'-','')
--get the access right id for full control select @rightID = (Select AccessRightID from ptsAccessRights where Name = 'Write') if(@rightID IS NOT NULL) --insert the records that are full access INSERT INTO ptsFunctionAccessEdges(FunctionAccessEdgeID, FunctionID, AccessRightID, UserID) VALUES(@strID, @arrayValue, @rightID, @userID) else RAISERROR(50100,15,1) END COMMIT TRAN tmp1 IF @@TRANCOUNT > 0 RAISERROR(50101,15,1) ROLLBACK TRAN tmp1
The transaction does not rollback when any errors occur at all. Any advice/help?
I am not an expert in Stored Procs. I would like to build one for a product list that would return a default value without using output parameters, if possible. Ultimately, I wouldn't be opposed to it.
It currently looks like this:
CREATE PROCEDURE ProductsByVendorNo ( @VendorNum nvarchar(24) ) AS
SELECT P.ProductID AS ProductID, P.ProductShortName AS ProductName, P.ProductDesc AS ProductDesc, U.UnitDesc AS Unit, P.VendorProductNumber AS VendorNumber, V.VendorName AS VendorName, P.Price AS Price, P.ImageThumb AS ImageThumb
FROM tblProducts AS P INNER JOIN tblUnitCodes AS U ON P.UnitCode=U.UnitCode INNER JOIN tblVendors AS V ON P.VendorID=V.VendorID
WHERE V.VendorsVendorNo = @VendorNum AND P.Inactive = 0
ORDER BY ProductName, VendorNumber
I would like for it to return a default, constant value for the URL in the ImageThumb field, if this one is empty. I could not find good documentation of how to use IF statements for this case, i.e. to alter the return of just one field.
I want to "deny" create, update,and delete access on the dbo stored procs that are in the database, but do not want take away dbo owner access. is this possible?
can i create a role and deny access on a particular table in msdb? or a system table in the user table. Thus preventing the developers on the box access to update any of the dbo owned sp's and have them create their own user-owned stored procs?
I am trying to set up a DTS package that selects data from one table on server A into another table on server B. I want to do a select statement that will do the following:
select store_name from store (server A) where date_created >= (select max(date_created) from store (server B)
I use EM 7.0 to manage all of my extracts, however, the data is moving from a Syabase (adaptive Server)onto another Syabase(adaptive server) machine. Unfortunately, there is no functionality for a linked server connection.
I tried the following, (which doesn't error out), but is not displaying the source columns in the destination tab of the DTS package when setting up the transformations.
declare @maxdate datetime exec serverB.dbo.sp_max_date_from_store @maxdate select store_name from store --(server A) where date_created >= @store
Any help or suggestions would be greatly appreciated! trevorb
I will be taking over a database that has almost no pk's or relations(this is not my choice, but a vendors) Management is looking at stored procs to improve performance, but I am wondering if the db is in this state will there really be a gain. I am pushing for normalization first, but if anybody has any ideas or opinions I would appreciate
