That is what appears to be happening. Only one pass of the table does all indexes.

>Did you happen to notice if it seemed to proceed any differently than it did before. What I have in mind is that, a little while back, someone made a (smart, I thought) suggestion that (re)indexing proceed doing only one pass of the subject table, building all of the required indices as it went along. It would be nice to try to figure out if something like that has been implemented, not to mention some tangible 'proof' that UT and wishes *do* actually get a listen.

>>I have experienced a drastic decrease in the time it takes to index/pack tables with SP3. I have a 1 million record table that took almost 30 minutes to index on 7 expressions. This now takes under 10 minutes. Has anyone else observed this. Wow!