Re: Disk space consumption: character varying(255) versus text usedfor index
От | Marcel Ruff |
---|---|
Тема | Re: Disk space consumption: character varying(255) versus text usedfor index |
Дата | |
Msg-id | 0edae0c2-528a-0bc6-fda1-942f1342e030@netwake.com обсуждение исходный текст |
Ответ на | RE: Disk space consumption: character varying(255) versus text usedfor index (Stephen Froehlich <s.froehlich@cablelabs.com>) |
Список | pgsql-novice |
I love PostgreSQL,
Marcel
Just to make it a little more clear, PostgreSQL has a binary UUID type that should house your GUID’s perfectly efficiently. (It also provides a lot of other useful functionality and speed improvements):
https://www.postgresql.org/docs/12/datatype-uuid.html
From: Justin <zzzzz.graf@gmail.com>
Sent: Friday, December 20, 2019 12:06 PM
To: Tom Lane <tgl@sss.pgh.pa.us>
Cc: ruff@netwake.com; pgsql-novice@postgresql.org
Subject: Re: Disk space consumption: character varying(255) versus text used for index
Have question this states the GUID is the text Hex value form a UUID if so that 36bytes long vs integer that is 128bits if that is the case convert this GUID to UUID type
or am i missing something?
On Fri, Dec 20, 2019 at 1:55 PM Tom Lane <tgl@sss.pgh.pa.us> wrote:
Marcel Ruff <ruff@netwake.com> writes:
> my btree index for
> guid | character varying(255)
> is currently 6.9GB (almost all guid entries are of size 37bytes)
> If I would change the data type to
> guid | text
> would the index size be reduced (without hitting performance)?
Wouldn't make any difference at all.
regards, tom lane
NetwakeVision
Alte Owinger Straße 100
D-88662 Überlingen
Phone: +49 7551 309372
http://www.netwakevision.com
http://www.royal-gps.com
В списке pgsql-novice по дате отправления: