I'm just TERRIBLY worried about links across tables breaking. I really appreciate your queries and for sharing them with us.
My fear with using them though is that I may get a collision, and then have a half-processed query or something.
I just think that if we all work togethe,r maybe we can use some more advanced SQl to not only combine some of those statements, but to do checking and stuff to make sure the numbers are being changed to unique values and stuff beforehand...
There used to be a fantastic graphic layout of the DB structure I remember seeing somewhere. Does anyone know if there is a current one?
I'm MORE THAN willing to help (perhaps work with you?) to come up with some queries for public consumption that can help people renumber their databases (in other words, "pack" and make sense out of" the IDs).
I just feel like I am flying blind.. because no matter how hard I try to understand the db structure (and I do to an extent), I'm ALWAYS afraid I'm going to miss something.
Another thing that we may wish to think about, consider, or even address... I wonder if there is a way to approach the way these things are created in the first place.
I'd certainly be willing to trade a BIT of performance for a more sane way of creating these IDs in the first place (maybe some code that allows you to override these values upon creation IF you want to... or even some code that searches and tries to create the IDs intelligently in the zone you're in.. or whatever)...
__________________
--
Keelyeh
Owner, ServerOp and Developer
Jest 4 Server
Linux (Jest3 runs on Fedora, our Dev servers usually run on Ubuntu and/or Gentoo), OC-12 Connection = Hella Fast
|