PostgreSQL Connection

Settings

Host	The host name (DNS name) or the IP address of the server on which the database is located.
Port	The TCP/IP port number used by the database.
User name	User name for the connection to the database.
Password	Password for the connection to the database.
Database	Name, schema, or instance of the relational database.
Fetch mode	For a PostgreSQL database, you can adjust the fetch mode as follows: Buffered (default): the query result is returned with buffering. In this case, no parallel queries to the same PostgreSQL Server are possible. Full: PostgreSQL returns the query result in one block. This is potentially slower, with higher memory consumption. It should only be set if the connection is used in several components that are processed in parallel during a load, e.g. several PostgreSQL extracts that are joined in a transform, an extract, and a load on the same PostgreSQL connection or when using parallel jobs. (Note that in those cases, if the fetch mode is not set to "full" this might lead to an error message: "postgresql portal <x> does not exist".)
Additional JDBC parameters	Optional: Database-specific parameters which are added to the generated JDBC URL.

For performance reasons, loading to a Postgres database uses (internally) the PostgreSQL command COPY, using streams for the source data (CopyManager API). To use the generic SQL update logic instead, this streaming backend can be deactivated by setting parameter #backend to value "gen" in additional JDBC parameters. For Postgres databases with version 10 or lower, this might be necessary in some cases.

For loading to PostgreSQL database the JDBC parameter #tcpKeepAlive may have to be set to "true" if a firewall is used, which closes the IP channel after a certain duration of inactivity. Otherwise the error message "An I/O error occurred while sending to the backend" may occur during the load.

Updated June 5, 2023