ROWID pseudocolumn (https://docs.oracle.com/cd/B19306_01/server.102/b14200/pseudocolumns008.htm#i1006255) helps to identify a row (not foolproof but its something). Would be great to have this added to the Kafka message in the 'source' struct (to help with homemade 'completeness' job that reconciles back to the oracle table. ie check that are at least all current ROWIDs in oracle db are showing in kafka. Especially useful for tables without PKs, as all cols defined in debezium as making the key).

For reference, some other oracle log miner projects getting this column:

https://github.com/averemee-si/oracdc/commit/c8356d6c4b53f93f89532cbdf4a8c9850407e500

https://github.com/erdemcer/kafka-connect-oracle/blob/0abb180cfd62f7878b8cb16f1aac27e1a8afc123/src/main/java/com/ecer/kafka/connect/oracle/LogMinerThread.java#L149

https://github.com/erdemcer/kafka-connect-oracle/blob/master/src/main/java/com/ecer/kafka/connect/oracle/OracleSourceTask.java#L236

https://venkyrao.medium.com/kafka-oracle-dcn-5703c4ca495f