The Dodgy GO Statement

  • Subramanyam Krishnamurthy-120784

    SSC Enthusiast

    Points: 123

    Comments posted to this topic are about the content posted at http://www.sqlservercentral.com/columnists/sKrishnamurthy/thedodgygostatement.asp


    Kindest Regards,

    Mani

  • colin naylor

    SSC-Addicted

    Points: 449

    The author seems to think that begin and end statements seem to do something in stored procedures where they do not.

    You can put a begin and end statement around whatever you like, but it wont stop the batch from continuing on to what comes next.

    You also don't even need a GO statement in order to create a stored procedure. The authors seems to have totally missed the purpose of the GO statement.

    Did you know, that SQL Server doesn't even understand a GO statement. It is not transact SQL!

    Hope this helps.

  • Alan Harley

    SSC Enthusiast

    Points: 160

    Do you really understand what happens when you create a stored procedure?

    Interesting question!  I agree with the above poster.  The BEGIN and END statements are nothing to do with stored procedures but are simply used to logically group statements in IF's/ELSE's, WHILE's and CASES's.

  • Fattman

    Old Hand

    Points: 330

    With you there.  There's more to GO than meets the eye, and it doesn't even have to be "GO" either!

    Thort it might be fun to share this early tale of GO horror (I was a lot younger then):

    SQL 7, asked to implement a script from a (VB) developer to add a column to a table on a remote, client mission-critical, really, really, really important customer address db. 

    So, did I double-check the code?  No. 

    Did I test it on a local db?  No. 

    Can you see a disaster coming?  Oh, yes...

    Run the script.  See the script run.  See the error.

    Turns out VB Fly Boy had scripted the change from Enterprise Mangler and it went something like this:

    SELECT * INTO newtable FROM oldtable

    GO

    DROP TABLE oldtable

    GO

    CREATE TABLE oldtable (with shiny new column)

    GO

    INSERT oldtable SELECT * FROM newtable

    GO

    Okey-dokey, except the first SELECT INTO failed.  But, thanks to the joys of "GO" the script ploughed on, dropped the table etc.  So, ended up with the new table with its new column and no data!

    Oops.  Well, that'd be ok 'cos the local dba on the ground will have a back up.  Well, yes, from 3 months ago.

    Lots and lots of lessons learned there.  I now spend my days doing macrame.

  • Peter Kryszak

    SSCarpal Tunnel

    Points: 4338

    I wonder if some of the confusion is from people transitioning from Oracle PL/SQL and other compiled languages where the language is much more sophisticated and clear.

    CREATE/BEGIN/END have clear syntactical and semantic roles in PL/SQL. In fact, the END statement includes the name of the procedure or function that is being defined.

    SUB/END SUB and FUNCTION/END FUNCTION in VB.NET are also very clear.

    T-SQL still seems like a toy language compared to others for this and some other reasons.

    --Peter

  • Mike C

    SSC-Insane

    Points: 23224

    "GO" is not a real SQL statement. It's a batch termination indicator for Query Analyzer. You might want to point out that trying to use "GO" in your scripts outside of Query Analyzer (like in a .NET SqlCommand), you can expect to catch some exceptions.

  • colin naylor

    SSC-Addicted

    Points: 449

    I think I already said that GO wasn't a real sql statement in my earlier message.

    Are we surprised that we haven't heard from the author?

  • sheepoo

    SSC Eights!

    Points: 959

    I am having real trouble understanding where I would use this 'approach' in real-world scenarios. In my experience with SQL Server (almost 2 years now) I have not come across one situation where the GO keyword caused me trouble. I use the SQLIDE Pro from Imceda/QUEST and the only time it uses the GO keyword is to do

    SET QUOTED_IDENTIFIER ON

    GO

    SET ANSI_NULLS ON

    GO

    at the start and

    GO

    SET QUOTED_IDENTIFIER OFF

    GO

    SET ANSI_NULLS ON

    GO

    at the end: I never had to use the GO keyword within a stored procedure.

    Maybe I am missing something here

  • colin naylor

    SSC-Addicted

    Points: 449

    No, you're not missing anything. You don't have to use the GO word when creating stored procs.

     

  • sheepoo

    SSC Eights!

    Points: 959

    Colin, thanks for the backup.

    That was exactly my point: You can live without GO as far as creating Stored Procedures are concerned (this being the point discussed in the article)

  • JT Lovell

    Ten Centuries

    Points: 1138

    Wow, tough crowd today.  Must be a Monday thing since yesterday was a day off for many folks.

    I use GO all the time in scripts, especially migration scripts as they have a mix of functions, stored procs, and SQL statements including drop and create statements many of which must be at the start of a batch.  Knowing where to put the GOs in the script make all the difference.

    One thing about GO that wasn't apparant to me when I first started TSQL was that you couldn't use a local variable that was declared "across a GO".  For instance this script gives you an error:

    --

    DECLARE @myvar varchar(100)

    GO

    SET @myvar = 'Bob was here!'

    SELECT @myvar

    --

    In such a short script it's easy to fix, but in larger scripts that need to process in a specific order, this limitation may require using temporary tables or other objects to hold values from one batch to the next.

     

  • Mike C

    SSC-Insane

    Points: 23224

    Yes, you mentioned that it's not a real Transact-SQL statement. I was expanding on what you said it "isn't" by explaining what it actually "is".

    As JT noted above, a discussion of how the QA "GO" batch terminator affects scope would be a nice addition to this article.

    Another nice addition would be the effect that "GO" has on scripts passed to SQL Server via other methods (such as .NET's SqlCommand).

    And to answer the author's question (just in case it wasn't fully evident in the other posts):

    Q: "...I don't understand why they allowed any executable statements after the end of stored procedure?"

    A: "Because the statements after QA's 'GO' are a different batch."

    Personally I think the author picked a good subject, but there are a lot of aspects he skipped over that would have made the article a lot more informative.

  • Simon Sabin

    SSCrazy Eights

    Points: 8142

    A couple of points, just to confirm that GO isn't TSQL but a terminator used by DMO and SMO to parse batches of SQL. Whats more with SQLCMD you can put a number after the GO for the preceeding TSQL batch to be repeated.

    The biggest issue I have seen is the deployment of sprocs. The easiest way is to concatenate all the files for the store procs together and run that combined file. However if one file doesn't have a go at the end and all the files start with a if exists drop statement you can end up with.

    If exists (select 1 ....)

        drop proc myfirstProc

    go

    create proc myfirstProc

        begin

        --    do some stuff

        end

     

    If exists (select 1 ....)

        drop proc mysecondProc

    go

    create proc mysecondProc

        begin

        -- Do some stuff

        end

    go

    If exists (select 1 ....)

        drop proc mythirdProc

    go

    create proc mythirdProc

        begin

        -- Do some stuff

        end 

    What you find is that this is likely to run in, however the first time someone runs myFirstProc it will drop mysecondProc Not good.

    On final note for those running scripts on SQL 2005 its a must that you start the script with :on error exit. This will avoid the problem detailed above, where one statement fails and but all the others are still run


    Simon Sabin
    SQL Server MVP

    http://sqlblogcasts.com/blogs/simons

  • Subramanyam Krishnamurthy

    Valued Member

    Points: 55

    I work in IST time zone. Now I am back to work.

    The intention in the article is clear, to highlight the unknown error/issues and its outcome with absence/usage of GO statement.

    Thanks guys for your valuable posts.

    Cheers,

    Krishnamurthy

     


    Mani

  • Sean Nolan-296153

    SSC Journeyman

    Points: 82

    All you have to do is check out Books Online

    GO

    Signals the end of a batch of Transact-SQL statements to the Microsoft® SQL Server™ utilities.

    Syntax

    GO

    Remarks

    GO is not a Transact-SQL statement; it is a command recognized by the osql and isql utilities and SQL Query Analyzer.

    SQL Server utilities interpret GO as a signal that they should send the current batch of Transact-SQL statements to SQL Server. The current batch of statements is composed of all statements entered since the last GO, or since the start of the ad hoc session or script if this is the first GO. SQL Query Analyzer and the osql and isql command prompt utilities implement GO differently.

Viewing 15 posts - 1 through 15 (of 20 total)

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