recently, the RowID value of the first row coming out of a DB Reader Node changed (in Knime 2.9.0-2.9.2). It has always been 'Row0', but since 2 days it is 'Row1' now, for all DB reader Nodes we use. I didn't install or upgrade anything, neither in the Database nor in Knime (It is a production server). This broke some of our production workflows. What might have caused that problem? Is it generally not safe to rely on the fact that the first row in a Knime dataset has the RowID 'Row0'?
Thanks in advance and best regards,
Thanks for your feedback. I can confirm that we haven't changed anything on the way we generate RowIDs in the Database Reader node since 2.8. The procedure calls ResultSet#getRow, if this fails we use our internal counter. Did you recently updated your driver? One other think that just came to my mind, with KNIME 2.9 we have added a couple of default database drivers which are part of the KNIME Common JDBC Feature. You might want to deinstall this feature under Help > About KNIME > Installation Details to make sure your own driver is picked up. Please let me know if this solves the problem?