I've had some fun in the past where I've had to grant a lot of tables and
other similar system commands. Unfortunatly, you can't use queries to fill
in fields for you. Anyway, I've implemented a patch which allows the
following:
grant select on ":2" to ":1"
\for select usename, relname from pg_catalog.pg_user, pg_catalog.pg_class where relname not like 'pg_%';
Produces:
Expanded to: grant select on "test" to "postgres"
Expanded to: grant select on "test2" to "postgres"
[etc]
Anyway, I'm sure you can come up with all sorts of uses. It occurs to me to
be a way to produce a whole lot of commands to create very similar triggers
based on a list of tables. And other commands to remove them again. My
questions are as follows:
1. Due to the way that backslash commands are implemented, the entire second
query has to be on one line. I don't see a good way to deal with this.
2. If the results being substituted contain quotes, they might not
substitute cleanly. Do we care?
3. Do people even want something like this?
4. Should it list the commands being executed?
5. What should happen upon an error in the generated query?
6. Maybe \foreach is clearer?
7. There should probably be a test mode to just dump the queries without
executing.
Thoughts welcome.
Preliminary patch at:
http://svana.org/kleptog/pgsql/forloop1.patch
Anonymous CVS access isn't working for me so it's not against the latest,
hopefully it's not too changed. It's somewhere around 7.3.2.
Have a nice day,
--
Martijn van Oosterhout <kleptog@svana.org> http://svana.org/kleptog/
> "All that is needed for the forces of evil to triumph is for enough good
> men to do nothing." - Edmond Burke
> "The penalty good people pay for not being interested in politics is to be
> governed by people worse than themselves." - Plato
