this surprises me.  Are you sure you are checking ERRORLEVEL correctly?

can you post the relevant part of your batch script? 

This is an ok script for some purposes perhaps, but it's missing a lot.  There are many options in CREATE INDEX that you are not covering, as you can see here:

don't be sorry - it was a valid question

this proc creates fine on my system: standard 2005 sp1

this free tool will script out all objects in any 2000 or 2005 database, using SMO. source is available so you can modify it as you please - it...

was the object you generated a script from in a 2000 or 2005 db?  if 2005, then what you are trying to do is not supported.

You might want to check out this free tool.  It's a cmd line app so it would be easy to schedule it to run nightly.  Also the source is available...

what sort of perf improvement do you expect from this change?  The only difference is the n-types take twice as much space. 

The only perf improvement I can see is...

Damian, I think you may be wrong there. 

This fails on my 2005 sp1 server:

create table #sp_help_job 
(
   job_id uniqueidentifier 
   ,originating_server nvarchar(30)     COLLATE Latin1_General_CI_AS 
   ,name sysname                        COLLATE Latin1_General_CI_AS 
  ...

if you are in control of all the sprocs involved, then there are many ways around this problem. 

however if you can't change the sproc you are calling with insert-exec,...

I am pretty sure this will fail on 2005 also.

insert exec will fail if the proc you are exec'ing is itself also using insert exec.

One kludgy way around it is...

well, ok. as you can tell I am a bit paranoid when it comes to dynamic sql.

If the user is just selecting what columns to select on...

fyi, Peter Larsson has implemented Djikstra in sql here:

http://www.sqlteam.com/forums/topic.asp?TOPIC_ID=77262

I wouldn't call this safe.

anyone that can control the process that inserts data into @tblIdFields can do damage to your server.

how does data get into this table variable?...

in that case have a look at sp_executesql in the article I linked to.  also it's documented in BOL. 

In any case Erland's article is a must read for anyone cosidering...