That's what I was looking for! Thanks!

the.hank, this is what I typically have done in the past if I run into a customer with your situation:

1) Verify the free space in each file (code listed below)

2)...

So you literally ran a DROP TABLE statement on the original table, see that it's gone from sys.objects, and when you try to import into that original table name again...

May not be the best way to do it, but you could create a stored procedure that accepts a starting ID and a number of records as input parameters and...

I can't find the site reference to the one I used when writing this but here's a code snippet from something I use to set up new servers using dynamic...

I agree with Jeffrey except that if you have very large databases and not exactly a lot of disk space, you may want to do the backups before just in...

Another handy light-weight tool you could use to find out what's locking the file is Unlocker found here: http://ccollomb.free.fr/unlocker/

I'm creating a new topic for this issue with a correct title. It will be called "CnstIsNotTrusted Returns True on Key Set with NOT_FOR_REPLICATION"

It's not intermittent meaning that sometimes we run the SELECT * INTO and it's slow and other times we run the same SELECT * INTO it's fast...

It's more like some...

It's really a SELECT * statement.

After further research we're seeing that the performance issues can occur using either a SELECT * INTO or an INSERT INTO depending on the situation...

Here is the code sample (replaced table names and key names with bogus names)

This causes CnstIsNotTrusted to be 1

ALTER TABLE [dbo].[TableA] ADD

CONSTRAINT [FK__TableA__TableB__ReportNameS] FOREIGN KEY ([ReportNameS]) REFERENCES [dbo].[TableB] ([ReportNameS])

ON DELETE...

I made a mistake, it seems to be the NOT FOR REPLICATION clause that's doing this.

I found that dropping / re-adding the foreign key constraint will reset the bit.

If there is a better solution please let me know.

I would create one maintenance plan for re-index and a separate one for update stats. Then schedule them when needed.

That should do it.

If you want to write your own proc then you can program it to only re-index if the logical frag and scan density are above certain thresholds (maybe 15% fragmented...