Passthru queries are literally just 'passed thru', to be executed as SQL on the source DB and they do NOT have access to the target DB. However any row returned will be subject to whatever transformation rules defined within Replicate including target_lookup.
Using just Replicate you cannot stop a source row from being processed by Replicate based on a target value, but you can stop it from being forwarded to the target.
Maybe you can use a DBLINK or other inter-database connection on the source DB itself to become part of the record selection expression in the source passthru query string? Then you could use target data in the source selection. Or maybe you can move (using replicate?) a target table from the target to the source just before full-load to help define which rows should, or should not be (re)loaded.
Hein.
