GO statement in SQL scripts

I have inherited a large script which uses DateFrom and DateTo parameters in 4 separate places in the script. I would like to DECLARE these parameters and put them into 2 variables to use in the 4 instances (instead of having to alter the script every time I run it). My problem is that the GO statement ends a batch of commands and the script loses the value of my variables.

Can anyone tell me exactly what the significance of the GO statement is in a SQL script, what it does and under what circumstances it is really necessary.

Thanks for your help.

itsmarkdavies@.hotmail.combasically you need a go any time you issue a:

CREATE DEFAULT, CREATE FUNCTION, CREATE PROCEDURE, CREATE RULE, CREATE TRIGGER, or CREATE VIEW statment. Also a table can not be altered and then referenced in the same batch.

Check out Books On Line, on the index tab type in the keyword "GO". BOL has a very goo explination of the use of "GO" and batches in general.|||So Paul it sounds like 'GO' is like 'commit' when you create new objects in your script ?|||Not quite, transactions are used to group operations into a logical work unit which you commit or rollback based on some test. A "go" is syntacticaly required to execute some TSQL statments. Consider the following:

Code:
----------------------------
begin transaction
create procedure test_getstuff as select db_name(), user_id()
exec test_getstuff
rollback transaction
if object_id('test_getstuff') is not null
print 'test_getstuff exists'
else
print 'test_getstuff does not exist'
----------------------------
will generate an error, however

Code:
----------------------------
begin transaction
go
create procedure test_getstuff as select db_name(), user_id()
exec test_getstuff
rollback transaction
if object_id('test_getstuff') is not null
print 'test_getstuff exists'
else
print 'test_getstuff does not exist'
go
sp_helptext test_getstuff
go
----------------------------
will work because TSQL requires a create procedure to be the first command in a batch. Notice, however, that the stored procedure does a little more than expected. Lastly:

Code:
----------------------------
begin transaction
go
create procedure test_getstuff as select db_name(), user_id()
go
exec test_getstuff
sp_helptext test_getstuff
rollback transaction
if object_id('test_getstuff') is not null
print 'test_getstuff exists'
else
print 'test_getstuff does not exist'
go
----------------------------
The 2nd go is used to terminate the batch. Now the sp looks correct.

Notice that in all of this I destroyed all my work by wrapping everything in a transaction. I could just ahve easily saved all my work by using a commit.|||Hi,

Use the "GO" statement if you want to end the SQL Statements and want SQL Server to execute till that.

Mr. Paul and Mr.Youngman has explained in a nice way.

You can add these below basic info also along with that.

Basically SQL Server must understand where Your SQL script ends.
So, Microsoft uses "GO" to terminate/end SQL Statements. Oracle uses ";" to terminate/end SQL Statements.

As soon as SQL Server sees "Go" statement it understands that SQL Statement is ended and it can start executing the SQL Statements till that point.

I can put in this way, like a man looking for pebbles. as soon as he finds a pebble, he picks and puts in a bag and looks nearby whether there are some more and keep going. Similarly SQL Server looks for "Go" and as soon as it sees "GO" it thinks, Ok I found a "go" and I have to end there and pick all the SQL Statements I come across and start executing.

Thats it.

Have Fun :)
Varad01