In an effort to optimise a load, I took all of the processing out and put in into a QVW that gets run before it to generate a QVD, so that in this QVW I can just do a load * from that QVD rather than doing load X as Y for all the fields. I went to test it, but when I reload the file it finishes, and says complete, and reaches the last command, but does not actually 'finish' it just sits there. If I try and do a limited load of only 1, it still gets to the end and then just sits there saying finished. Interestingly the title bar says 'waiting', and at this point no CPU is being used, but as soon as I click on 'close' it changes to say 'finished' and one of the four CPU's goes to 100% usage. I can't for the life of me work out whats going on, as even if I comment out all of the updated code, it still just sits there?
I had this happen to me once. I wasn't able to recover. I had to redo my changes starting from an earlier version of the document that I'd saved. I'm guessing that somehow the file had gotten corrupted.
In my experience, hanging at the end of the load is a problem with the way the tables are joining together. But when you do a limited load of one record, even problematic table structures should complete just fine. And it sounds like this is just a single table anyway.
I'm really hoping someone has a better idea than "your file is corrupt; start over", but that's all I have.