Keith Fiske

Yes, the value itself is a supported thing in PostgreSQL. But does that table have any constraint boundaries on the `event_ts` column?

So looking into this more, I see how you're expecting this to work. However, as you saw, an infinite value cannot be partitioned out. By definition `infinity` is greater than...

Perhaps I can make it a flag to the partition_data procedure in 5.x. But just to be sure, infinite values are considered valid in your case and this wasn't errant...

Ok. I'll see if I can make this an option for the partition_data functions in the future. I'd recommend working with NULL if that's a possibility, though

Apologies, NULL isn't going to work either. I currently have a requirement that the partitioning column be NOT NULL. Also this would run into the same issue. Which table would...

Yes, I can see that. I'll see about adding that option in with a caveat in the docs that it can affect maintenance performance.

Thanks for the suggestion. I'll try and look into it when I have a chance.

@james-johnston-thumbtack Thank you for the investigation on this. You might want to also bring this up on the official PostgreSQL mailing lists as a missing inheritance feature for native partitioning....

I have a beta PR up that should automatically inherit the REPLICA IDENTITY from the parent table. If you're able to test, any feedback would be appreciated https://github.com/pgpartman/pg_partman/pull/615

Version 5.1 has been released with this fix