From: | John DeSoi <desoi(at)pgedit(dot)com> |
---|---|
To: | Ivan Sergio Borgonovo <mail(at)webthatworks(dot)it> |
Cc: | pgsql-general(at)postgresql(dot)org |
Subject: | Re: looping over a small record set over and over in a function |
Date: | 2009-06-20 13:54:31 |
Message-ID: | D02B9D8F-D55D-4C66-BFA3-79C716882790@pgedit.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-general |
On Jun 20, 2009, at 8:35 AM, Ivan Sergio Borgonovo wrote:
>> And lastly, if your current approach really is the only way to
>> compute what you're after, then maybe PL/pgSQL isn't the right
>> match for the problem; it looks like you'd be better served by a
>
> Yeah. I gave a look to python but I don't want to add one more
> language to the mix.
> I enjoy strict type checking of plpgsql even if some bit of
> syntactic sugar would help to make it more pleasing and I think it
> is the most lightweight among the offer.
> Still I don't know how easy it is with eg. python to load an array
> with a result set, change it and place it back into the table where
> it was coming from.
>
>> language that can work with arrays of typed structures. As I'm not
>> familiar with the other PL languages I can't tell whether they
>> would be suitable in that respect, but I suspect Python or Java
>> would be able to handle this better.
>
>
> Your suggestion about cursor could be the way... but I don't know
> enough about cursors internals to understand if updating a field of
> a cursor will cause disk writes.
I have not tried this, but the documentation says arrays can be
created for "any built-in or user-defined base type, enum type, or
composite type." So maybe you could define a composite type and stuff
those into a single array?
John DeSoi, Ph.D.
From | Date | Subject | |
---|---|---|---|
Next Message | benoît carpentier | 2009-06-20 14:09:17 | Benetl, a free ETL tool for files using postgreSQL, is out in version 2.8 |
Previous Message | Hartman, Matthew | 2009-06-20 13:19:59 | Re: Dynamic table |