Partitioning in PostgreSQL is a long-awaited functional. Although Postgres can implement partitioning through inheritance, this approach has a number of drawbacks, such as the need to manually create sections and support triggers, significant overhead costs for scheduling, and lack of runtime optimizations. In the report, we will talk about the extension of pg_pathman, which we are working at.
Pg_pathman supports HASH and RANGE partitioning and performs optimization in the planning and execution stages, supports quick insertion by eliminating triggers in favor of Custom Node, contains functions for managing sections (add, split, merge, etc.), FDW-based sharding, non-blocking data migration and other possibilities.
We will also tell about integrating the pg_pathman into PostgresPro Enterprise Edition and supporting Oracle-like syntax for partitioning. In conclusion, we will tell about the new partitioning capabilities in PostgreSQL 10, which is already implemented and the ways of further development.