Re: Extended customizing, SQL functions,

pgsql@mohawksoft.com

From: pgsql@mohawksoft.com
To: "Jonathan Gardner" <jgardner@jonathangardner.net>
Cc: pgsql-hackers@postgresql.org
Date: 2004-05-29T03:54:37Z
Lists: pgsql-hackers
> On Friday 28 May 2004 04:08 pm, pgsql@mohawksoft.com wrote:
>>
>> No transactions.
>>
>
> Then it won't work. Let's say we take your example. The variable "s" is
> supposed to be the sum of a column in a table.
>
> When someone modifies the data in the table, when is "s" adjusted? What if
> it is modified but the transaction is aborted? Who sees the "correct"
> value
> of "s" at any point in time?

Like I said, it is a *bad* idea. No one is denying this, it "breaks" all
the rules, but "with care" I can see a limited, but valuable function.

>
> --
> Jonathan Gardner
> jgardner@jonathangardner.net
>