>Hello,
>
>We have a vfp desktop application that has been running for many years without any problems. In the last month or so, we are experiencing problems with corrupt indexes. Primarily it has been restricted to one table, although I believe one other table had its index corrupted as well. This is now occurring about every other day.
>
>Another thing is it is not trashing the entire file, but it is basically mis-placing one record. Once it occurs and someone tries to search that record, they get the corrupted index message. We can scan throught the table using the index and it will find every record except that one that has been "corrupted".
>
>After we rebuild the index, everything runs fine.
>
>We are checking everything we can think of, including network cards, network lines, switches, etc. Best we can tell, it does seem to be happening on computer(s) from one corner of the building, although we can not be absolutely sure.
>
>If anyone has any suggestions, we are open to trying anything at this point.

If by "rebuilding" the index, you mean you're just issuing REINDEX, that may not be enough to get rid of some types of corruption. What I'd recommend:

1. Find a known good copy of the .CDX file - from backup, your dev machine, whatever. Doesn't matter how old it is or if it's up-to-date

2. Make sure no-one is using the production table

3. Overwrite the production .CDX with your known good one

4. Then, REINDEX

Another thing to check is if any field in the bad row (or nearby row(s)) has bad characters such as CHR( 0 ) (which appears as a vertical black bar in a BROWSE). I've seen those cause index problems.