Keith Fiske

I haven't had the time to address this issue yet. I have a large refactor planned for later this year or the first part of next year that I can...

Will lift this restriction once trigger-based partitioning has been removed in 5.0. Not sure if LIST partitioning support will be added yet at that time.

Just to clarify, you are saying there are no child tables at all in this partition set?

Ok. I wanted to make sure. pg_partman cannot work without having at least one child table. Otherwise it has no idea what the "next" table to create is. The create_parent()...

The issue here is that you're not going to be able to run `create_parent()` if you're already restoring the contents of the `part_config` table. You'll get a duplicate key violation...

Another option would be to use the `dump_partitioned_table_definition()` on the original system to generate the `create_parent()` call along with an `UPDATE` statement to restore the remaining configuration options. Note this...

Just checking to see if any of the suggestions i made worked for you? I won't be merging this PR at this time.

Sorry for the late response on this, but took a while to sit down and play around with this. So, I don't think PostgreSQL supports primary keys and self-referential foreign...

Work colleague (@pgguru) pointed out an alternative approach here. You can detach the child table first, then drop it. ``` keith@keith=# ALTER TABLE data_mart.events DETACH PARTITION data_mart.events_p2022_07_02; ALTER TABLE Time:...

Created a new label to mark this as an upstream issue. I'm not seeing any way to be able to handle this in pg_partman at this time. Unless you have...