Most likely you've hit a point where the stats need to be maintained manually and not count on the server to do this.

The server waits for 20% row change in...

roger.plowman (5/25/2011)

---

So to sum it up, this is about understanding what the user wants, having the time to implement it, test it, and prepare for the inevitable scaling and utter...

You need to do a cross apply for the 12 months of the period to make sure you don't have nulls in that column.

You can also consider having 1 client...

Do you have nulls in the date or month somewhere in the data?

What's the event? Can you provide a quick screenshot?

In what column do you see the prepared...

How is the data inserted / updated?

If the application generates the next PK (insert) and then does an update and fills up all the rest of the columns you're going...

Insert into rpttable ()

Select from prod where not exists (check in rpttable)

Reported to Steve... spacing is wrong. The article is complete. but very weirdly spaced.

HTH :w00t:.

SUM(CASE WHEN TranType = 2 THEN Qty ELSE 0 END) AS QtyReturns

, SUM(CASE WHEN TranType = 1 THEN Qty ELSE 0 END) AS QtySold

Whisper9999 (5/24/2011)

---

I'm using a pretty old version. I will try the latest version to see if it runs faster. Plus, now it has an exclusion table. So...

#1 turn auto_shrink OFF and reindex the whole db.

#2 select * from diarioliq where dataliq = '2011-05-10' bad perfomance

That is likely because the stats are out of date.. literally. ...

You mean you can't figure out out to split the sales and returns???

Something like this maybe :

Qty * Amount * CASE WHEN TransactionTypeCode = 2 THEN -1 ELSE...

I use this one and works perfectly for me :

http://sqlfool.com/2010/04/index-defrag-script-v4-0

What scan / sample mode are you using?