Again, this is because my procedure also removes gaps in the sequence, hence why it updates all the records. If you try my procedure with different parameters and look at the cache hits it will always hit, whereas the other one will recompile for different parameters combinations. As I stated in an earlier post, I will remove the resequencing to remove gaps and restate at zero into the delete routine, enabling my procedure to also work on a range and drop drastically in complexity. Apologies if you have not read back this far. My procedure suffered by being to all-encapsulating, as if you had a broken sequence due to deletes of say 1,5,7,11,13, mine would set this back to 0,1,2,3,4, whereas the others wouldn't - hence why it is complex and hence why it updates all records.

Cheers, james