Larry,

Hmmm, I've been cathed by my own words... How bad can it get ? ;-) Well, of course in very extreme circumstances it might be a good idea but even with 9 million deleted records (of 20 million) It is likely to be more than a burden than a help. But o.k. never say never... just look into the entrire situation and above all, test it in a real world situation (thus in a loaded network if it is going to be used on one).

Thanks for the correction,

Walter,

>Walter,
>Never say never! While not likely, if the table had 9 million deleted records, a DELETED() tag may be a good idea. As you have said in the past, look at the situation in its entirety.
>
>Regards.
>
>>When dealing with a 20 million record table you never want an index on deleted(). Even if it contains a million deleted records; when doing any query or just using the table when set delete = ON, it downloads 19 million indexkeys from the network, which could take forever, especially when the network bandwith is limited. Another discriminating (selective) index would be far more efficient.
>>
>>Walter,