...
This step allows you to load data into a Vertica database table. It uses the VerticaCopyStream class in the Vertica JDBC driver (documented on the Vertica site in the Programmer's Guide > Vertica Client Libraries). Using the VerticaCopyStream class with a COPY statement is the highest performance method offered by Vertica to stream bulk data into a table.
Note: This step creates an internal buffer and allocates memory for chunks of 1000 rows by default if the row size allows this. You need to ensure enough memory to the JVM to avoid Out Of Memory exceptions. We recommend to allocate at least 2GB heap memory for the JVM.
Prerequisites
- The Vertica JDBC driver must be in the PDI classpath.
- You must be connecting to a Vertica cluster version 5 or higher. Both Vertica CE and Vertica EE are supported.
- The host machine running the transformation using this step must be able to connect to all nodes of the Vertica cluster via TCP.
- The Vertica user configured in the connection must either be a superuser, or have USAGE privilege on the schema and INSERT privilege on the table.
...