Hi,

If you're planning to segregate the tables by month, I'm not sure why you think there will be a problem with your original primary key of [ItemDate],[ItemKey] ?

If you pass...

Vinu,

Could you post some code as well please?

I'm guessing here, but I suspect that it may be to do with statistics (the values that let the optimizer decide how many...

Just doing one myself Martin. What exactly is it that you're having trouble with?

Colin - thanks for that. I confess I had actually found that reference on Monday, but thanks for pointing it out anyway! (I meant to put a message...

It uses histogram statistics. These are samples of the data throughout the table at regular intervals. The values at each sample enable the optimizer to get a view...

Hi Bill,

We're only talking about one system here. At the risk of repeating myself, this problem only exists for one database - all other databases ON THE SAME...

Thanks Antares - sorry if I seemed to rant slightly, but I just seemed to be going over the same ground several times which was getting frustrating.

I am already planning...

Thank you, and that goes for everybody, for taking the trouble to contribute to this thread. However, I do wish people would read it completely.

Yes, I KNOW that the...

Well, it's always worth pointing such things out because we all forget the obvious at times. But no, I have checked that (several times!). It's also worth pointing...

Hi Scott,

I'll try it (doesn't do any harm as you say), but I doubt it was ever replicated. Even if it had been, I wouldn't be able to truncate...

Just to revisit this a little - I am intending to try Tom's suggestion, since I suspect this is a corrupt logfile.

However, if I issue a manual CHECKPOINT command, the...

Thanks for all the replies, but I should clarify that ai am NOT talking about SHRINKING the log - just truncating the active portion. The problem is that if...

But the log ISN'T truncating at checkpoint. The recovery interval is set to default, but even when I leave my test transactions (and there are lots of these) to...

Best combination to use is a BACKUP LOG ... command followed by a DBCC SHRINKFILE command.

BACKUP LOG [ with truncate_only ]

DBCC SHRINKFILE ()

If the log records...

Actually, unless I've misread some of the other replies, I think the concensus is (just) to defrag first, then rebuild indexes.

I've never found any formal documentation to support this, but...