PostgreSQL 10 High Performance - Third Edition

List partitioning

The example here, and the pattern you're most likely to deploy, uses range partitioning. This is where you provide a non-overlapping range of values that each partition uniquely includes. It's also possible to partition based on an explicit list of values that direct which partition a value goes into. For example, if you had queries against the customer's table that were routinely targeted at specific states, you might partition based on that:

CHECK ( state IN ( 'AK', 'AL', 'AR','AZ' ))
CHECK ( state = 'CA' )  

So, that populous states like California (CA) have their own partition.

Note that whether a partition is a range or a list is a choice of descriptive terminology. There's no syntax difference between the two, and to ...