Re: [HACKERS] INSERT INTO view means what exactly? - Mailing list pgsql-hackers

From Tom Lane
Subject Re: [HACKERS] INSERT INTO view means what exactly?
Date
Msg-id 3852.927658408@sss.pgh.pa.us
Whole thread Raw
In response to Re: [HACKERS] INSERT INTO view means what exactly?  (jwieck@debis.com (Jan Wieck))
Responses Re: [HACKERS] INSERT INTO view means what exactly?  (jwieck@debis.com (Jan Wieck))
List pgsql-hackers
jwieck@debis.com (Jan Wieck) writes:
> Tom Lane wrote:
>> regression=> INSERT INTO z VALUES ('bar');
>> INSERT 411636 1
>> 
>> OK, where'd tuple 411636 go?

>     Tuple 411636 went into data/base/regression/x :-)

.../z, you meant --- yup, I see you are right.  Weird.  I didn't
realize that views had an underlying table.

>     It  might be a good idea to abort if there's a SELECT rule on
>     the result relation but not  one  for  the  actual  operation
>     performed.  I'll  put  that  onto  my personal TODO for after
>     v6.5.

I agree, that would be a good safety feature.
        regards, tom lane


pgsql-hackers by date:

Previous
From: Tom Lane
Date:
Subject: Re: [HACKERS] create index updates nrows statistics
Next
From: Tom Lane
Date:
Subject: Re: [HACKERS] pg_dump core dump, upgrading from 6.5b1 to 5/24 snapshot