Comment 11 for bug 1236979

Hmm, I hadn't looked at depesz comments since I first started looking at this. I think I've basically come full circle and put the inefficient stuff into my suggestion.

I will try and test this on some data here.

If the cursors are what is speeding this up, then we might be able to do the original query as a cursor and forget all the looping.

If the joining is what is slowing it down then we will have to loop over just the copy data.

I will try with retrieving just the new items per circ lib via a cursor and new items for all libs via cursor.