Chris-475469 (7/24/2013)

---

I just wanted to know that, assuming the fact table would be archived by date, would have something like 3 partitions (Current month, current year, history) sound like the...

Rob-350472 (7/24/2013)

---

I've added the OPTIMIZE hint to the end of the procedure - I can't quite believe this, results sets from the proc and query are IDENTICAL as a result....

DECLARE @d varchar(6) = '20114'

SELECT try_convert(date, substring(@d, 1, 4) +

...

Rob-350472 (2013-07-23)

---

It's almost as if the SP is running with different permissions to the straight code or something. What I don't get, if that's the case why wouldn't it throw...

I looked at the thread and the plans, and I like to throw my hat into the ring, and suggest a possibility that has not been mentioned yet: a bug...

Then you apparently have "go" where it should not be.

The reason your initial attempt failed, is because SQL Server compiles the entire...

I would add an INSERT trigger that captures app_name(), host_name(), SYSTEM_USER, USER and sysdatetime() to a log table.

Today, and it has been this way for a decade or more, applicaitons typically connect, run a query, disconnect and rely on the connection pooling the API that a reconnection...

And what date do you expect 20114 to convert to? And by which rule?

That's great to hear. Note that the article suggests other alternatives that are faster than the simple version you tried.

It would help if you could share the text for the first job step (which I assume is the one that gets hung).

If you want a suggested out of the...

Could you explain more clearly what you did which you wish you hadn't down? What did you download and install by mistake? You say "error". Does this mean that you...

I think that you are correct in that BULK INSERT is more conservative on memory than a list-to-table function.

As for the setup as such, I guess SSIS fans gringe, but...

Yes, there are some differences in the Setup program. The Setup for SQL 2008 is generally better, although you need to be careful with the collation, because it is shoved...

UPDATE TableA

SET col = CASE WHEN b.cnt >= 1

THEN...