Re: Strange inconsistency using psql - Mailing list pgsql-admin

From John Scalia
Subject Re: Strange inconsistency using psql
Date
Msg-id 8C7A5C93-FBA2-49A7-AF59-1314585F67A0@gmail.com
Whole thread Raw
In response to Re: Strange inconsistency using psql  (Keith <keith@keithf4.com>)
Responses Re: Strange inconsistency using psql
List pgsql-admin
Correct, but my question really is, why is VIEW different from all the other types of objects? Shouldn’t the word VIEW also be an optional qualifier for the name of the object? Inconsistency strikes me as an oversight.

Sent from my iPad

On Dec 18, 2019, at 12:41 PM, Keith <keith@keithf4.com> wrote:




On Wed, Dec 18, 2019 at 12:29 PM John Scalia <jayknowsunix@gmail.com> wrote:
I had to do some privilege assignments this morning on a bunch of tables, sequences, functions, and views. On all of these object, I generally try to use a command like:

GRANT all ON TABLE x TO new_user;

Where TABLE is either that object or a SEQUENCE or FUNCTION. These all worked perfectly for me. What did not work was specifying that the object was a VIEW. The system would spit out a syntax error at the object’s name being specified, however, if I omitted the word VIEW, and not specify the type of object, then the GRANT succeeded.

Was this intentional behavior, or is the grammar slightly amiss? All the other types of objects worked perfectly with this style of command.

Jay

Sent from my iPad


There is no VIEW clause to the GRANT command. The TABLE clause is actually optional in the command when setting privileges on tables or views. And generally you can use the privileges used on tables when setting privileges on views.


Keith

pgsql-admin by date:

Previous
From: Keith
Date:
Subject: Re: Strange inconsistency using psql
Next
From: Pepe TD Vo
Date:
Subject: backup script error with could not connect to database