I hope you guys don't mind me jumping in here but I've been reading the debate on the index on DELETED() and it intrigued me enough to try it on an order entry system I've got running. I didn't have an index on DELETED() on any of the tables until I tried testing a few days ago. I put the index in about 12 tables that range from about 500 records (300k) to a little over 180,000 records (22MB). Some of the tables have only a few (<1%) deleted records and others have what I consider alot (~20%). The system is run on a small network of about 7 full-time users. They complained about speed problems (constantly!) and I thought I'd try this index to see if it would help. I just got back this afternoon from their site and they said it did make a difference but it wasn't a really big one. I know this is not very scientific but it does represent an end-user reaction of at least this application of the index. Frankly, I didn't see much of a difference in my own limited testing here on my development machine and some of their databases. Most LOCATES and SEEKs only netted me about a 3 hundredths of a second speed increase and some didn't do anything that I could really notice. Nevertheless, if the users are seeing SOME improvements (and no noticable slowness in record inserts/updates!) then I'm happy with the easy solution.

- A Hilton