From: | Michael Lewis <mlewis(at)entrata(dot)com> |
---|---|
To: | Alvaro Herrera <alvherre(at)2ndquadrant(dot)com> |
Cc: | Dennis Ryan <dennisr1963(at)outlook(dot)com>, pgsql-general <pgsql-general(at)postgresql(dot)org> |
Subject: | Re: Hash partitioning, what function is used to compute the hash? |
Date: | 2020-05-12 15:51:27 |
Message-ID: | CAHOFxGq6uX7=PNqkbpC5yptnGgtHJi-MmaTNjBDJSDxZsUEwKA@mail.gmail.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-general |
On Mon, May 11, 2020 at 3:13 PM Alvaro Herrera <alvherre(at)2ndquadrant(dot)com>
wrote:
> On 2020-May-11, Michael Lewis wrote:
>
> > Afaik, hash partition doesn't have real world expected use cases just
> yet.
>
> I don't think I agree with this assertion.
>
I didn't mean to be critical at all, or even make a statement of fact. Just
sharing my impression. I typically view partitioning from the perspective
of multi-tenancy and with the restrictions on primary keys & partition
keys, I can't typically use partitioning except for audit logging tables
and then range partitions make the most sense there because of doing
backups and dropping the oldest data. Perhaps it is just that hash has
never been the right tool for my use cases. I'd love to know some real life
examples of when hash partitioning was the best option.
From | Date | Subject | |
---|---|---|---|
Next Message | Tom Lane | 2020-05-12 15:54:40 | Re: ESQL/C: a ROLLBACK rolls back a COMMITED transaction |
Previous Message | github kran | 2020-05-12 15:40:25 | Re: AutoVacuum and growing transaction XID's |