we run ASA8.0.2.4339

we've got a scenario where we are dumping the contents of
certain tables from the Consolidated database into text
files of the format 'insert into table_name...'. we are
then using the generic 'read' command at the remote to read
the insert statements. we *expect* many of these to fail
because the Remote database already contains those records,
thus we have ON_ERROR = 'off'...the problem for us is that
even when these inserts fail with a 'UNIQUE' (primary key)
at the Remote database, it seems that upon its next
synchronization that it attempts to send even these failed
inserts up to the Consolidated(where they then fail again,
obviously)..

is there a way to prevent failed statements at the Remote
database from going into the log file so the do not
replicate? with perhaps a dbeng8/dbmlsync switch or database
variable?

TIA

Jon wrote:
I gotta say, I'd be VERY surprised if the behaviour you are describing
is actually happening. You are essentially saying that the engine is
logging a committed operation when that operation in fact fails.

I recommend you try building a very simple repro to prove that this is
(not) happening.

g.f
--
Greg Fenton
Random Dude
--------
Visit the iAnywhere Solutions Developer Community
Whitepapers, TechDocs, Downloads
http://www.ianywhere.com/developer/