Dynamic table partitioning in postgres

Question

I was looking up ways to have postgres partition data into tables based on timestamp for example, but without having to add the relevant child tables manually. I saw this blog post that does just that

https://blog.engineyard.com/2013/scaling-postgresql-performance-table-partitioning

but I'm dubious about the idea of creating tables based on string concatenation and checking the pg_catalog. Is this a reasonable idea?

Answer 1

pg_partman is an extension created specifically to manage the complexity of partition management. I haven't used this extension, but I've used others by the same author and they are generally of excellent quality.