To quote my reply from above:

If there is more than that on your side...

I would use a CASE statement to test for Call_BeginTime being a numeric value or a datetime value and do the math based on that.

Here's an example using variables:

DECLARE

@Call_BeginDate...

I cannot verify what you describe.

Since you didn't provide any sample data supporting your statement I made some up.

DECLARE @t TABLE (id int, data xml)

INSERT INTO @t

SELECT 1,'

<Block>

...

bigclick (12/23/2009)

---

Our vendor was able to give us an example.

Call Begin Date: 12/03/2009

Call Begin Time: -100

Call End Date: 12/04/2009

Call End Time: ...

oliver.morris (12/23/2009)

---

brief thought could you use a view to pull in the dynamic data and run the pivot t-sql on the view - you would need to now what the...

First of all: Don't loop! (unless you'll get paid per second your code runs...) 😉

You could use UNPIVOT (see BOL for details) to get the table "flattened".

With the resulting table...

Is there any column available to know which rows belong together (even an ID column would be ok as long as all rows after one entry with 'email' would belong...

Same...

Well, then your friend must have had a lot more information available than we did...

How else could he come up with Row_Number when there was nothing in your post indicating...

Well, I doesn't look like the results for those two variables will always be NULL...

Reason: the query most probably won't run at all.

I just tested the part with the subselect...

Why concatenate?

Since both are int, I'd use

SELECT Year * 100 + Period From TableA.

You could use CTEs (common table expressions) to determine @Total2 (see BOL for details).

This would make the whole query much more readable and most probably would make it more obvious...

What exactly are you trying to do?

The way this proc is written it will fail as soon as you have more than one row having [ReportId]=@ReportId.

Even though you added ORDER...

The physical order doesn't really matter that much...

If your table has a clustered index, new rows will be inserted in the physical order, maybe causing to physically move other rows...

What have you tried so far?

If nothing yet you might want to look into BETWEEN clause in BOL (BooksOnLine, the SQL Server help system).