• When working with PostgreSQL, you can use the "ctid" pseudo columns instead of primary keys
  for row identification. See #36
• Under rare circumstances SQL statements were generated that could not be processed by Sybase database systems.
  This has been corrected.
• The performance of the working table scope "local database" was improved. Especially with embedded usage (API).
• If the subsetter is used via API, access to the working tables is synchronized
  so that different extraction models can be used concurrently.