From: | andrew <andrew(dot)ylzhou(at)gmail(dot)com> |
---|---|
To: | pgsql-hackers(at)postgresql(dot)org |
Subject: | Re: look up tables while parsing queries |
Date: | 2006-02-06 09:24:26 |
Message-ID: | 29d3b0b90602060124u62ef5c85p3176b5435a1d4f66@mail.gmail.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
On 2/5/06, Neil Conway <neilc(at)samurai(dot)com> wrote:
> If you're referring to the raw parser (parser/gram.y), you should not
> attempt to access any tables. For one thing, the raw parser might be
> invoked outside a transaction. The statement might also refer to a table
> created earlier in the same query string, which would mean the
> referenced table would not exist when the latter part of the query
> string is parsed.
>
> Instead, database access should be done in the analysis phase -- see
> transformStmt() in parser/analyze.c and friends. There are plenty of
> examples in the code of how to access tables, which should be a helpful
> guide.
>
> -Neil
It is not in the raw parser. I meant inside the transformStmt(). What
is "friends"? Could you possibly point out at least one place that
illustrates how to access tables?
Thanks.
--
andrew
From | Date | Subject | |
---|---|---|---|
Next Message | Pavel Stehule | 2006-02-06 09:58:43 | slow information schema with thausand users, seq.scan pg_authid |
Previous Message | Mark Woodward | 2006-02-06 05:17:27 | Re: Shared memory and memory context question |