>I think there is a big difference between simply having a TAG on DELETED(), vs a TAG on DELETED ** FOR DELETED() **; ie. the former indexes ALL records, while the latter only indexes "deleted" records.
>
>I use DELETED() tags all over the place, since I do record recycling, and haven't really noticed any problems ...

Hi
That's because filtered indexes are not used in Rushmore optimisations.

I've made some tests in the past and I came to the conclusion that indexes with low selectivity can make queries much slower than without them, specifically on low bandwith networks (we had 10M network at the time) and tables with many records.

If is about a table with millions of records, a query that selects 500 of them will be much faster without a tag on DELETED(), with SET DELETED OFF and the HAVING !DELETED() clause in query, no matter if the query is "fully optimisable" in both cases.

Dorin