Re: BUG #3002: PQexecParams only supports some commands; needs improved error reporting, documenting or fixing - Mailing list pgsql-bugs

From Tom Lane
Subject Re: BUG #3002: PQexecParams only supports some commands; needs improved error reporting, documenting or fixing
Date
Msg-id 19059.1171438759@sss.pgh.pa.us
Whole thread Raw
In response to BUG #3002: PQexecParams only supports some commands; needs improved error reporting, documenting or fixing  ("Phil Endecott" <spam_from_postgresql_bugs@chezphil.org>)
Responses Re: BUG #3002: PQexecParams only supports some commands; needs improved error reporting, documenting or fixing
List pgsql-bugs
"Phil Endecott" <spam_from_postgresql_bugs@chezphil.org> writes:
> I understand that PQexecParams may only be used for SELECT, INSERT, DELETE
> and UPDATE commands, but not for other commands where parameter substitution
> might be useful such as CREATE VIEW and CREATE TABLE AS SELECT.  When I used
> PQexecParams with CREATE VIEW with one $ parameter I got no error
> immediately, but got a "no value found for parameter 1" error when I later
> ran a query that used the view.

As near as I can tell, your issue is that you haven't thought carefully
about what "$1" in a CREATE VIEW *means*.  Is it supposed to represent a
constant value bound into the view when it's created?  Is it supposed to
represent a value supplied whenever the view is used (if so, how do you
disambiguate when the query calling the view uses $1 for its own
purposes?).  Or something else?

Given defensible arguments for a particular answer to these definitional
questions, we might think about implementing something, but right now my
instinct is to not go there.  I'm not seeing a use-case that would
justify work on this.

            regards, tom lane

pgsql-bugs by date:

Previous
From: Tom Lane
Date:
Subject: Re: BUG #3000: Linking problem
Next
From: Tom Lane
Date:
Subject: Re: BUG #2996: to_char( timestamp, 'DD-Mon-YYYY HH24:MI:SS.MS' ) reports .1000 ms