Re: Backend-internal SPI operations - Mailing list pgsql-hackers

From Tom Lane
Subject Re: Backend-internal SPI operations
Date
Msg-id 29343.967846134@sss.pgh.pa.us
Whole thread Raw
In response to Re: Backend-internal SPI operations  (Jan Wieck <janwieck@Yahoo.com>)
Responses Re: Backend-internal SPI operations
List pgsql-hackers
Jan Wieck <janwieck@Yahoo.com> writes:
>     Hmm  -  too  simple  - real life is harder. So to what do you
>     expand the query

>         SELECT a, c, d FROM my_view, other_table
>             WHERE my_view.a = other_table.a
>             AND other_table.x = 'foo';
       SELECT a, c, d    FROM (SELECT a, b, c FROM my_table) AS my_view, other_table           WHERE my_view.a =
other_table.a          AND other_table.x = 'foo';
 

I'm still not detecting a problem here ... if selecting from a view
*doesn't* act exactly like a sub-SELECT, it'd be broken IMHO.

We're not that far away from being able to do this, and it looks more
attractive to work on that than to hack the rewriter into an even
greater state of unintelligibility ...
        regards, tom lane


pgsql-hackers by date:

Previous
From: Tom Lane
Date:
Subject: Re: Backend-internal SPI operations
Next
From: Jan Wieck
Date:
Subject: Re: Backend-internal SPI operations