From: | brian <brian(at)zijn-digital(dot)com> |
---|---|
To: | pgsql-general(at)postgresql(dot)org |
Subject: | Re: Calling Python functions with parameters |
Date: | 2008-07-21 16:45:53 |
Message-ID: | 4884BD41.1060808@zijn-digital.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-general |
Jerry McRae wrote:
> I am having a problem with the simplest of Python functions, so I must be
> doing something wrong. After hours of searching and trying many options, I
> need the key that my puny brain is missing here.
>
> I cannot pass parameters to a plpythonu function. I have tried within psql
> and with pgAdmin III (which adds IN or INOUT to the parameter list - which I
> can't find documented). I'm an advanced Python programmer but a beginning
> PostgreSQL user.
>
> Here is what I have, which I copied almost verbatim from example code:
> ----------------------------
> test_dev-# \p
> create or replace function testf5i(a integer,b integer)
> RETURNS integer AS $$
> if a > b:
> return a
> return b
> $$ language plpythonu
>
>
> test_dev-# \g
> CREATE FUNCTION
> test_dev=# select testf5i(1,2);
> ERROR: plpython: function "testf5i" failed
> DETAIL: exceptions.NameError: global name 'a' is not defined
That looks like the same error I received in my first attempts at
plpython. After muddling with it for some time (and this was my first
practical Python code at all, so you can imagine how confused I was), I
added (the equivalent to your code) at the beginning:
a = a
b = b
As ridiculous as that looks, it's what worked for me. Although, function
contained a class with several methods (the function creates "slugs" on
the fly from proper names for use in URLs, eg. Élisabeth Carrière ->
elisabeth_carriere). So, maybe my own problems were related to that.
PlPython remains a mystery to me, in many respects. It's certainly *not*
the best way to learn Python.
> PS. What I need to do, which I also could find not examples on the mailing
> lists or the Internet, is to de-normalize some tables (user, addresses,
> phones, emails) into one big view and then update the proper tables upon
> updates. The web application then can just get one row and not have to deal
> with all the different tables
>
> ...
>
> (And I'm surprised that I could find no examples of anyone already doing
> this? Is there something inherently wrong with this approach?)
Perhaps it's simply not worthwhile. That seems like a great deal of work
just to avoid having to deal with several tables. (I didn't end up using
my slug-creator function, myself, and moved that logic into the
application scripts.)
b
From | Date | Subject | |
---|---|---|---|
Next Message | btober | 2008-07-21 17:04:15 | Re: How to remove duplicate lines but save one of the lines? |
Previous Message | Scott Marlowe | 2008-07-21 16:29:10 | Re: How to remove duplicate lines but save one of the lines? |