Capturing The Error Description In A Stored Procedure

  • Amit Jethva

    SSCarpal Tunnel

    Points: 4707

    Comments posted to this topic are about the content posted at http://www.sqlservercentral.


    -- Amit



    "There is no 'patch' for stupidity."



    Download the Updated SQL Server 2005 Books Online.

  • Peter E. Kierstead

    SSCarpal Tunnel

    Points: 4114

    To catch those elusive error messages you could also try:

    Declare @SQL nVarChar(4000)

    Set @SQL=N'osql -Usa -P -Q"declare @t table(PKey Int not Null) insert @t select Null"'

    Exec xp_CmdShell @SQL

    and do whatever you want with the response



    PeteK
    I have CDO. It's like OCD but all the letters are in alphabetical order... as they should be.

  • jbomma

    SSC Enthusiast

    Points: 152

    Amit, excellent posting!!!!

    I am in same kind of situation.. by setting Xact_abort off you are loosing transaction control.

    BTW, I could not able to find your code snippet.

    Can you please forward to me at my email.... jagadeeswar.bomma@gmail.com

    Thanks


    jay

  • Bill Nye

    Hall of Fame

    Points: 3188

    Another alternative follows. Note that the error message contains details such as

    the parameter values at the time of error and the name of the stored procedure

    (or you could refer to a section of SP). You might also consider adding details

    like the userid.

    The point of this is to show that you CAN get error details out of a stored procedure,

    even though the default return value is only an INT datatype. The example shows a

    raiserror and a print of the error message - in some cases you may want to raise the

    error within a stored procedure, while at other times you may want to handle it

    outside the SP, such as the print statement that could have just as easily handed

    the error message to your calling code or to a logging process.

    [font="Courier New"]

    IF EXISTS (SELECT * FROM dbo.sysobjects

    WHERE id = OBJECT_ID(N'[dbo].[usp_errorhandling]')

    AND OBJECTPROPERTY(id,N'IsProcedure') = 1)

    DROP PROCEDURE [dbo].[usp_errorhandling]

    GO

    SET ANSI_NULLS ON

    GO

    SET QUOTED_IDENTIFIER ON

    GO

    --------------------------------------------------------------------

    --Purpose: Error handling example

    --Changes:

    -- Bill Nye 4/29/08 Created

    --------------------------------------------------------------------

    CREATE PROCEDURE [dbo].[usp_errorhandling]

    @Branch int ,

    @Division int ,

    @errormsg varchar(1000) output

    AS

    SET NOCOUNT ON

    DECLARE @errorid int

    Select 1/0 -- Purposely cause error

    Set @errorid=@@Error

    IF @errorid <> 0

    BEGIN

    SELECT @errormsg=description FROM master.dbo.sysmessages

    WHERE error=@errorid

    SET @errormsg='Error updating when Division_ID='

    + Cast(@Division as char(1)) + ' and Branch_ID='

    + Cast(@Branch as varchar(3)) + ' in usp_errorhandling: '

    + @errormsg

    RAISERROR(@errormsg,16,1)

    END

    GO

    -------------------------------

    DECLARE @RC Int, @errormsg varchar(1000)

    EXEC @RC = [dbo].[usp_errorhandling] 999, 9, @errormsg OUTPUT

    Print @errormsg

    Print @rc

    GO

    DROP PROCEDURE [dbo].[usp_errorhandling] -- Cleanup[/font]

  • tmahmood

    SSC Enthusiast

    Points: 101

    Hi Amit

    I tried to download your code attached with this topic, But it doesn't work.

    Will you please check it and email me your code, that will be great.

    tahirmd@hotmail.com

    Regards

  • tmahmood

    SSC Enthusiast

    Points: 101

    Thanks Amit

    Using your Idea I'm able to solve "Capturing The Error Description In A Stored..." for SQL 2000.

    I would like to share my investigation with you guys.

    Here are my findings:-

    With this query I'm trying to create duplicate records in pubs.dbo.authors. By calling SP I get SQL error message.

    USE [PUBS]

    GO

    INSERT INTO authors

    values('172-32-1176','Doe','Jone','408-496-7223','Bigge Rd.','Menlo Park','CA','94345',1)

    DECLARE @retval INT

    EXEC usp_Get_ErrorMessage @retval output

    Results:

    [errNumber: 2627] [errState: 1] [errLevel: 14] [errInstance: MyLocal] [errLine: 1] [errProcedure: NULL] [errMessage: Violation of PRIMARY KEY constraint 'UPKCL_auidind'. Cannot insert duplicate key in object 'authors'.]

    (1 row(s) affected)

    Note:- Store procedure "usp_Get_ErrorMessage" is attached with this thread. rename .txt with .sql

    Best of Luck

    Logician --> [Tahir]

  • Lasse Schioettz

    Right there with Babe

    Points: 739

    Bill,

    Many of the sysmessage strings contain placeholders that should be fed variables run-time. Extracting the string only will produce messages like this:

    "Cannot insert the value NULL into column '%.*ls', table '%.*ls'; column does not allow nulls. %ls fails."

    Essential information about which column and which table is missing.

    Sincerely,

    Lasse

  • Ryan Henning

    Valued Member

    Points: 54

    Thanks Pete!

Viewing 8 posts - 1 through 8 (of 8 total)

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