From: | Robert Haas <robertmhaas(at)gmail(dot)com> |
---|---|
To: | Peter Geoghegan <pg(at)heroku(dot)com> |
Cc: | Pg Hackers <pgsql-hackers(at)postgresql(dot)org> |
Subject: | Re: Are we sufficiently clear that jsonb containment is nested? |
Date: | 2015-06-26 15:05:27 |
Message-ID: | CA+TgmoYRugo7EBD_kstGNC2yVDwHsJjqYfS852Aq+2wL8zJdDQ@mail.gmail.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
On Wed, Jun 24, 2015 at 5:53 PM, Peter Geoghegan <pg(at)heroku(dot)com> wrote:
> I worry that "8.14.3. jsonb Containment and Existence" is not
> sufficiently clear in explaining that jsonb containment is nested.
> I've seen anecdata suggesting that this is unclear to users. We do
> say:
>
> """
> The general principle is that the contained object must match the
> containing object as to structure and data contents, possibly after
> discarding some non-matching array elements or object key/value pairs
> from the containing object.
> """
>
> I think that we could still do with an example showing *nested*
> containment, where many non-matching elements/pairs at each of several
> nesting levels are discarded. This could be back-patched to 9.4.
> Something roughly like the delicious sample data, where queries like
> the following are possible and useful:
I would be fine with adding a *compact* example of this kind to the
table that begins section 8.14.3. I probably would not back-patch it,
because the absence of that example is not an error in the
documentation, but I will not complain if someone else does.
--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
From | Date | Subject | |
---|---|---|---|
Next Message | Robert Haas | 2015-06-26 15:10:34 | Re: WIP: Enhanced ALTER OPERATOR |
Previous Message | Tom Lane | 2015-06-26 14:54:31 | Re: Nitpicking: unnecessary NULL-pointer check in pg_upgrade's controldata.c |