Weird scenario involving SQL and native data management

["James H. H. Lampert" <jamesl@xxxxxxxxxxxxxxxxx>]

Has anybody ever heard of anything like this scenario?

User creates an SQL view of a file, with a select on a specific field. 
We don't know whether the physical file originated natively, or was 
itself created through SQL.

Looking at this logical through native Data Management, it initially 
shows an extra record, with the field on which the select was made 
blank. At this point, we don't even know whether this extra record was 
an actual record from the physical file, or came out of thin air.

Then the user opens the file in DFU. That's about the only think 
specific we know about the scenario: that they looked at it, but did not 
update any records, in DFU.

After that, and forevermore, the file behaves properly when opened 
through native Data Management.

Anybody ever hear of anything like that? Or why using DFU to merely 
inspect a logical generated as an SQL view would correct the file's 
behavior?

--
JHHL