tableupdate() in transaction

[JimPowers]

Have a batch posting program that adds records to a number of tables. Tables to be updated are using optimistic table buffering with multilocks set on. When all records have been added, changes are committed with tableupdate(.T.,.F.) inside a transaction. If any tableupdate() return .F., execute a ROLLBACK and a tablerevert() on all tables, otherwise go to END TRANSACTION. I've run into a situation where one of the tables does not contain the new records at the end of the transaction, but all of the others do. In debug, I verified that tableupdate() returned .T. for all of the tables. This code has been in use for a long time with no problems. Any ideas of what set of circumstances could be causing it to break now?

 

[perryf]

Have you rebuilt the indexes on the table in question?

PF

 

[fluteplr]

Are you sure that table actually got updated in the first place?

If you do a tableupdate on a table that had not had its data changed it, I believe it will still return .t.

 

[JimPowers]

Yes, the records were added. Stepping through in debug, I saw where the APPEND BLANK's and REPLACE's were executed. Also, the RECCOUNT() of the table increased by 1 each time a record was added, then went back to its original value at END TRANSACTION.