pg.ddx.io  pgsql-sql@postgresql.org mailing list archive  
help / color / mirror / Atom feed
how to convert relational column to array?
3+ messages / 2 participants
[nested] [flat]

* how to convert relational column to array?
@ 2005-12-19 16:06 george young <gry@ll.mit.edu>
  2005-12-19 16:54 ` Re: how to convert relational column to array? Michael Fuhr <mike@fuhr.org>
  0 siblings, 1 reply; 3+ messages in thread

From: george young @ 2005-12-19 16:06 UTC (permalink / raw)
  To: pgsql-sql

[PostgreSQL 8.1.0 on i686-pc-linux-gnu, compiled by GCC gcc (GCC) 4.0.1]
I'm trying to convert a column from a traditional relational form to an array:

create table old_tab(name text, id int, permits text);

newschm3=# select * from old_tab order by name;
   name   |  id   |  permits   
----------+-------+------------
 baker    |   581 | operator
 lawless  |   509 | operator
 lawless  |   509 | originator
 lcalvet  |   622 | originator
 loomis   |   514 | operator
 loomis   |   514 | originator
 pig      |   614 | operator
 pig      |   614 | originator
 pig      |   614 | supervisor

create table new_tab(name text, id int, permits text[]);

-- I insert one row per name:
insert into new_tab select distinct name,id,cast('{}' as text[]) from old_tab;

Now I want to fold all the 'permits' values into the new permits arrays.
I can do:

update new_tab set permits=new_tab.permits||ot.permits from old_tab ot where ot.name=new_tab.name and ot.permits!=all(new_tab.permits);

but this only gets one permits value per name.  Repeating this many times would
eventually get all of them, but it seems there must be a more reliable way?
[I don't care about the *order* of permits values in the array, since order did not exist in old_tab]

Just to be clear, I want to end up with:

newschm3=# select * from new_tab order by name;
  name   | id  |                 permits                  
---------+-----+------------------------------------------
 baker   | 581 | {operator}
 lawless | 509 | {operator,originator}
 lcalvet | 622 | {originator}
 loomis  | 514 | {operator,originator}
 pig     | 614 | {operator,originator,supervisor}

-- George Young
-- 
"Are the gods not just?"  "Oh no, child.
What would become of us if they were?" (CSL)



^ permalink  raw  reply  [nested|flat] 3+ messages in thread

* Re: how to convert relational column to array?
  2005-12-19 16:06 how to convert relational column to array? george young <gry@ll.mit.edu>
@ 2005-12-19 16:54 ` Michael Fuhr <mike@fuhr.org>
  2005-12-19 17:12   ` Re: how to convert relational column to array? george young <gry@ll.mit.edu>
  0 siblings, 1 reply; 3+ messages in thread

From: Michael Fuhr @ 2005-12-19 16:54 UTC (permalink / raw)
  To: george young <gry@ll.mit.edu>; +Cc: pgsql-sql

On Mon, Dec 19, 2005 at 11:06:12AM -0500, george young wrote:
> create table new_tab(name text, id int, permits text[]);
> 
> -- I insert one row per name:
> insert into new_tab select distinct name,id,cast('{}' as text[]) from old_tab;
> 
> Now I want to fold all the 'permits' values into the new permits arrays.

In PostgreSQL 7.4 and later you can build an array from a select,
so I think the following update should work (it did for me when I
tested it):

UPDATE new_tab SET permits = array(
  SELECT permits
  FROM old_tab
  WHERE old_tab.name = new_tab.name AND old_tab.id = new_tab.id
);

-- 
Michael Fuhr



^ permalink  raw  reply  [nested|flat] 3+ messages in thread

* Re: how to convert relational column to array?
  2005-12-19 16:06 how to convert relational column to array? george young <gry@ll.mit.edu>
  2005-12-19 16:54 ` Re: how to convert relational column to array? Michael Fuhr <mike@fuhr.org>
@ 2005-12-19 17:12   ` george young <gry@ll.mit.edu>
  0 siblings, 0 replies; 3+ messages in thread

From: george young @ 2005-12-19 17:12 UTC (permalink / raw)
  To: pgsql-sql

On Mon, 19 Dec 2005 09:54:49 -0700
Michael Fuhr <mike@fuhr.org> threw this fish to the penguins:

> On Mon, Dec 19, 2005 at 11:06:12AM -0500, george young wrote:
> > create table new_tab(name text, id int, permits text[]);
> > 
> > -- I insert one row per name:
> > insert into new_tab select distinct name,id,cast('{}' as text[]) from old_tab;
> > 
> > Now I want to fold all the 'permits' values into the new permits arrays.
> 
> In PostgreSQL 7.4 and later you can build an array from a select,
> so I think the following update should work (it did for me when I
> tested it):
> 
> UPDATE new_tab SET permits = array(
>   SELECT permits
>   FROM old_tab
>   WHERE old_tab.name = new_tab.name AND old_tab.id = new_tab.id
> );

That's exactly what I needed.  Works great.

Thanks,
	George
-- 
"Are the gods not just?"  "Oh no, child.
What would become of us if they were?" (CSL)



^ permalink  raw  reply  [nested|flat] 3+ messages in thread


end of thread, other threads:[~2005-12-19 17:12 UTC | newest]

Thread overview: 3+ messages (download: mbox mbox.gz follow: Atom feed)
-- links below jump to the message on this page --
2005-12-19 16:06 how to convert relational column to array? george young <gry@ll.mit.edu>
2005-12-19 16:54 ` Michael Fuhr <mike@fuhr.org>
2005-12-19 17:12   ` george young <gry@ll.mit.edu>

This inbox is served by DDX for PostgreSQL; see mirroring instructions
for how to clone and mirror all data and code used for this inbox