@danielbb
Can you provide a more detailed example of your use case? You can use a batch input to periodically read all rows, but your question implies a need to index newly inserted and/or updated rows.
Some DBMS products provide internal timestamp columns.
A batch input may be acceptable for small tables.
A scheduled search combining a key column lookup cache, dbxquery, a differencing operation, and collect may also be acceptable; however, a target table that contains duplicate foreign keys but no unique key column may be difficult to track.
