>Heya Jim!
>
>CDX bloat or no, and reindex speed aside, if the cdx header gets corrupted, you cannot reindex. So why try to use reindex? I just tried a reindex on a million records which took 13 seconds. There are three tags, field1, field2 and field3. The 3 index on commands took 11 seconds.

Hi Mike,

I thought I suggested why in the initial message.

I thought I also said that keeping INDEX ON statements around in case of corruption was a necessary thing regardless. But I do have to ask myself how often does a CDX header get corrupted.

As for the 13 secs. vs 11 secs, I've just gotta figure that cacheing played a role here.

Cheers,
JimN