I have a large unpivot task that seems to be unpivoting invalid data. I have 336 columns input columns, these are 24 sets of the same type of data. The primary key for the row is being passed through. The pivot key value is 1, 2, 3, 4.... 24 Running this through an unpivot there are rows appearing where there should not be any. The data in those rows seem to correspond to a row with a different primary key. When I reduce the number of columns being pivoted down to 96, still 24 sets everything works fine. This seemed to be the highest number I could get to before the problem appeared. Initially I though it may be a problem with Sql Server 2005 so updated to a trial of 2008 enterprise, then tried SP1 of 2008, then SP2 of 2008, then the March 2011 cumulative update for SP2 of 2008 and still the same errors. Has anyone encountered a similar problem?

No - but if you can replicate this behaviour in a simple package, please post it up to Connect. If you're able to provide that package, my experience is that they can usually comment on it within a few days. You can also let us know where that sample is, in order to get confirmation that lends more weight to your report. (Please post the link to the Connect submission back here if you do that.) Talk to me now on

connect won't let me register, justs throws errors back at me. Additional information: the error only occurs if I keep nulls from the data source.

Can you post up a simple package somewhere else for us to look at? Talk to me now on