The Applier reads only committed changes from the generated intermediate files and applies these changes to the target. The Applier combines all changes from committed transactions since the last apply operation and then applies these changes to the target in one transaction. After the data is committed, it is purged from the intermediate files.
Usually, you run the Applier on the target system. However, you can run the Applier on another system provided that it can access the target database by using the target database connectivity client or ODBC.
If an apply operation fails, the entire batch of changes is rolled back on the target database. If you restart the Applier, or if you shut down the database and then restart the database, the Applier can accumulate the changes and restart from the last checkpoint.
The Applier can apply source DDL changes to targets for which DDL replication is supported. DDL replication is not supported for Apache Kafka targets.