Hi Emmanuel and happy new year to you and your family,
The way metadata fetching is implemented in PostgreSQL's JDBC driver involves many complex joins on system tables. It happens to be very slow (you have the same problem if you need ResultSetMetadata and you don't enable Sequoia's MetadataCache). We should probably have a 'schema cache' to reduce the schema fetching time especially when a lot of tables or key constraints are enforced.
I worked a bit more on this problem. I filled an improvement request in JIRA and attached a candidate patch to fix the problem. See https://forge.continuent.org/jira/browse/SEQUOIA-885 for more information. -- Guillaume _______________________________________________ Sequoia mailing list [email protected] https://forge.continuent.org/mailman/listinfo/sequoia
