Re: [HACKERS] Refactor handling of database attributes betweenpg_dump and pg_dumpall
От | Haribabu Kommi |
---|---|
Тема | Re: [HACKERS] Refactor handling of database attributes betweenpg_dump and pg_dumpall |
Дата | |
Msg-id | CAJrrPGfit1_zOF7woh6MHUn2136ye71sDJFy6Xb=be+P+ROuoQ@mail.gmail.com обсуждение исходный текст |
Ответ на | Re: [HACKERS] Refactor handling of database attributes betweenpg_dump and pg_dumpall (Robert Haas <robertmhaas@gmail.com>) |
Ответы |
Re: [HACKERS] Refactor handling of database attributes betweenpg_dump and pg_dumpall
|
Список | pgsql-hackers |
On Thu, Oct 26, 2017 at 10:01 PM, Robert Haas <robertmhaas@gmail.com> wrote:
On Mon, Oct 23, 2017 at 7:36 AM, Haribabu Kommi
<kommi.haribabu@gmail.com> wrote:
> Apologies for not providing much details.
>
> pg_dumpall is used to produce the following statements for database,
>
> "Create database" (other than default database) or
> "Alter database set tablespace" for default database (if required)
>
> ACL queries related to database
> Alter database config
> Alter database role config
>
> whereas, pg_dump used to produce only "create database statement".
How about adding a new flag --set-db-properties that doesn't produce
CREATE DATABASE but does dump the other stuff? -C would dump both
CREATE DATABASE *and* the other stuff. Then you could dump built-in
databases with --set-db-properties and others with -C.
Thanks for the idea, Here I attached the patch that implements the same.
The newly added option is not recommended to be used in normal cases and
it is used only for upgrade utilities.
In case if user issues pg_dump with --set-db-properties option along with --create
or --clean options, an error is raised. Currently there is no way to throw an error
in case if the dump is generated with --set-db-properties and try to restore with
--clean option. To avoid this change, we may need to add additional details in the
archive handler, but is it really needed?
Regards,
Hari Babu
Fujitsu Australia
Вложения
В списке pgsql-hackers по дате отправления: