Re: Proposal "VACUUM SCHEMA"

* Alvaro Herrera (alvherre(at)2ndquadrant(dot)com) wrote:
> Multi-table CLUSTER uses multiple transactions, so this should not be an
> issue. That said, I don't think there's much point in CLUSTER SCHEMA,
> much less TRUNCATE SCHEMA. Do you normally organize your schemas so
> that there are some that contain only tables that need to be truncated
> together? That would be a strange use case.

I could see it happening in environments which use schemas when doing
partitioning. eg: data_2014 contains all of the data_201401-201412
monthly (or perhaps weekly) tables.

> Overall, this whole line of development seems like bloating the parse
> tables for little gain.

Still, I see this point also. I do think it'd be really great if we
could figure out a way to segregate these kinds of DDL / maintenance
commands from the normal select/insert/update/delete SQL parsing, such
that we could add more options, etc, to those longer running and less
frequent commands without impacting parse time for the high-volume
commands.

I'm less concerned about the memory impact, except to the extent that it
impacts throughput and performance.

Thanks,

Stephen