Hi Stephen, and no further benchmarking required, thanks. I was just wanting an idea of the differences and I appreciate that this is from the dim past.

I admit to flabbergastation at such numbers!

Thanks again (must do some thinking on this)

Jim N

>>Stephen,
>>
>>How much of a timing difference did you observe (in seconds, not in percentage) in OPENING??
>>
>>Regards,
>>
>>Jim N
>>
>>>>Stephen,
>>>>Yes, for recycling rows, you are using the favorable side of the skew.
>>>>I never checked opening times and I don't understand why it would be faster. I'll try to do some testing.
>>>>
>>>>>>Charlie
>>>
>>>I have no idea why there is a difference, but it helps in tables with a 10% or more deleted content and set deleted on.
>
>This is all old memories and a few beers since {grin}, but, 115 seconds for this :
>
>stime=seconds()
>use my_big_table
>etime = seconds()
>wait wind nowait str(etime-stime)
>brow
>
>Without deleted() tag it was close to 200 seconds.
>
>That table had 4 tags on necessary columns as well as an added on on deleted(). Total # of rows? 1.5 to 1.75 million. Had a 10%, 25%, and 40% reation for deleted. As the precentage grew, the opening time got worse with no deleted() tag.
>
>These were all ground zero tests, off of a Novell server with 125 meg of ram, 100mb nics, P200 machine with 96 meg ram and plenty of temp file space. That means shutdown, restart, relogin, restart the exe that did just the above.
>
>I'm under deadline pressure to re-do the test now butr maybe after Labor day weekend?
>
>Your reslults may be very different on NT servers, becasue of NT file open / closing differences.
>
>__Stephen