Re: JSON doc example (matchiness) - Mailing list pgsql-hackers

From Alexander Korotkov
Subject Re: JSON doc example (matchiness)
Date
Msg-id CAPpHfdsyHJtniiqo-tmfd74VEQpOzDAxHwS1Xh0sRH_-1msK7A@mail.gmail.com
Whole thread Raw
In response to Re: JSON doc example (matchiness)  (Alexander Korotkov <aekorotkov@gmail.com>)
Responses Re: JSON doc example (matchiness)
List pgsql-hackers
On Sun, May 9, 2021 at 4:01 AM Alexander Korotkov <aekorotkov@gmail.com> wrote:
> On Sat, May 8, 2021 at 7:09 PM Erik Rijkers <er@xs4all.nl> wrote:
> > On 5/8/21 3:48 AM, Michael Paquier wrote:
> > > On Fri, May 07, 2021 at 10:18:44PM +0200, Erik Rijkers wrote:
> > >> The JSON doc has this example (to show the need for double backslash):
> > >>
> > >> $ ? (@ like_regex "^\\d+$")
> > >>
> > >>
> > >> The example is not wrong exactly, and can be cast to jsonpath, but as-is can
> > >> never match anything.
> > >>
> > >> I think it'd be helpful to provide that example so that it more probably
> > >> matches when the user does a quick trial.
> > >>
> > >> Llet's change it to something like:
> > >>
> > >> $.* ? (@ like_regex "^\\d+$")
> > > Ah, I see.  What you are telling here is that we match the regex on
> > > the full JSON string, which is pretty useless, and you are suggesting
> > > to change things so as we'd match with the key names at the first
> > > level.  Makes sense.
> > >
> > > This paragraph of the docs say:
> > > "For example, to match strings that contain only digits"
> > > Could we be more precise here?  "strings" looks to much generic to
> > > me in this context when actually referring to a set of path of keys in
> > > a JSON blob.
> >
> > Yes, "string values"  is probably another small improvement.
>
> What about the attached patch?  Wording "string values of the root
> object" seems most precise to me.

I propose backpatching this to 12 when jsonpath was introduced.  It
seems useful to have this docs improvement every release supporting
jsonpath.

Objections?

------
Regards,
Alexander Korotkov



pgsql-hackers by date:

Previous
From: Tom Lane
Date:
Subject: Non-reproducible valgrind failure on HEAD
Next
From: Andres Freund
Date:
Subject: Re: Non-reproducible valgrind failure on HEAD