opc.three (6/21/2011)

---

I would recommend temp tables here given the expected number of rows (~215K). Te optimizer always assumes a rowcount of 1 for table variables so you have a better...

Lynn Pettis (6/20/2011)

---

Have you tried putting the results of the query using just the contains predicate into a temporary table (not a table variable), and then query the results? ...

@Lynn Pettis

Thanks for taking a look.

I actually haven't tried the query you suggested so I ran it against some live data just now.

The query adds an additional nested loop to...

@opc.three:

Thanks for the suggestion.

The CTE will actually have to be flipped to include the CONTAINs predicate.

The reason being that the CONTAINs predicate cannot be applied to a table that is...

Looks like we might have missed each other's reply.

If this wasn't the case, then I'd like to clarify: I did not create the TVF. It is executed by the...

@Lynn Pettis

The TVF is for the FullTextMatch.

Is it possible to actually retrieve the code SQLServer executes for these?

If it is, I would appreciate a short how-to.

The call is from the...

Our observations on the TVF pretty much match (on my first post).

Tests were executed on the same server, data, connection.

There's no difference in the execution plan from running TOP 1...

My apologies. I misunderstood what you meant by graphical format.

I've attached the *.sqlplan for both top 1 and top 2.

@opc.three:

Yep, that's the only difference in the actual execution plan. I mentioned the correlated parameters being different in my original post.

I've attached the images: Top1 and Top2. I hope this...