I have a front end system on client machines and the backend on the network.

Occasionally it becomes corrupted with  a record losing its autonumber
primary key in the main table. At times I am able to either do a compact and
repair on the backend, or I may have to delete the record and then do a
compact and repair. At other times the corrupted record wont delete and I
have to use an uncorrupted/record empty copy of the table and use a query to
copy all records from the corrupted table to the copy. In this case the
corrupted record will not copy to the new table and is effectively deleted.

I hadn't been able to work out why the corruption has been occurring, but I
suspected it was a locking issue. There are about 10 concurrent users.

I have been doing some further development and came across something which
triggered the problem.
The front end always has open a form which has as its recordsource the main
table. I was developing another form which when a particular field is
updated it uses a function to update a field in the main table.

The problem occurs (i.e. corrupts a record and then the table) if the record
being updated on the new form by the function happens to be open in the
hidden main form. (Although it happens intermittently). The  "Write
Conflict" dialog box comes up (i.e. "Save Record"  "Copy to clipboard"
"Drop Changes") and if either "Save Record"  or  "Drop Changes" is chosen
the record may become corrupt.

Any help greatly appreciated

John Baker