Hi You could try to read both extracts with osmosis and use a merge task before writing them both to the database in one rush. This may result in overlap at the borders, depending on the data sources. Peter