QA only stores the text of the query in the .sql file

ijaz raises a good point!

In case it's not what you're after (but it probably is and is a great solution), here is the answer to your question.

COALESCE(CAST(YourIntCol AS VARCHAR),'')...

I've come across this before and it was a surprising 'feature'

This is what I found on MSDN.

Whether SQL Server interprets an empty string as either a single space or as...

What if you formatted each output col like this:

 COALESCE(yourCol, '') + ','

Or am I being overly simplistic? let me know

Thanks!

That's pretty likely, especially since you say it's using the index now.

My calc was a rough guess for 4.4M rows (I may have misread your earlier post).

Let me clarify this: when it was taking 4-8 seconds, how many rows was it...

PS: Thank you for posting the DDL - All too often forgotten by posters

4 to 8 seconds to return 4.4M rows isn't all that surprising (that's pretty fast, actually). Your query has to transfer huge amounts of data to the client (800-1100+MB if...

Actually, that’s a good point Ronald. The original poster mentioned an acceptable range of numbers that is encapsulated in his UDF. It is possible—and therefore must be considered—that his...

I’ll say this first:

I’ve reread the thread and, out of the three solutions, your solution actually seems to solve the problem by creating whole random numbers that are equally...

I agree that memory consideration is a very important ase are hundreds of other issues that must be considered when desiging a database or writing TSQL. Perhaps my statement was...

It sounds like you're running SQL2005 Ninja. In 2005, I see pretty decent IO performance. Running the query in 2000 shows similar results, but scans go from 46 to 54...

My bad PW. I missed your comment about using master. It might be worth editing your post to qualify your object names, in case anybody makes the same mistake I...

One more thought,

If you are cross joining system tables, as one poster suggested, rather than creating the values in memory then consider using the NOLOCK hint. Also be aware...

By chance, does your table have an identity column?

What is the schema of this table?

And, because people are going to ask, can you tell us a little more about this...