Re: jsonb_object() seems to be buggy. jsonb_build_object() is good. - Mailing list pgsql-hackers

From Andrew Dunstan
Subject Re: jsonb_object() seems to be buggy. jsonb_build_object() is good.
Date
Msg-id e9d17417-0e10-5b17-d396-1b2214ca90b6@2ndQuadrant.com
Whole thread Raw
In response to Re: jsonb_object() seems to be buggy. jsonb_build_object() is good.  (Bryn Llewellyn <bryn@yugabyte.com>)
Responses Re: jsonb_object() seems to be buggy. jsonb_build_object() is good.
List pgsql-hackers
On 2/16/20 7:25 PM, Bryn Llewellyn wrote:
>
> B.t.w., you earlier said “The double quotes  [around “dog”] serve a specific purpose, to allow values containing
commasto be treated as a single value (see syntax details for the exact rules) in the resulting array of text values.”
Butthis test shows that they are not needed for that purpose:
 


I didn't say that. Someone else did.


>
> select jsonb_pretty(jsonb_object(
>   '{a, 17, b, dog house, c, true}'::varchar[]
>   ))
>
> This is the result:
>
>  {                    +
>      "a": "17",       +
>      "b": "dog house",+
>      "c": "true"      +
>  }
>
> The commas are sufficient separators.
>
> It seems to me, therefore, that writing the double quotes gives the wrong message: they make it look like you are
indeedspecifying a text value rather than a numeric or integer value. But we know that the double quotes do *not*
achievethis.
 
>


No, you haven't understood what they said. If the field value contains a
comma it needs to be quoted. But none of the fields in your example do.
If your field were "dog,house" instead of "dog house" it would need to
be quoted. This had nothing to do with json, BTW, it's simply from the
rules for array literals.


cheers


andrew


-- 
Andrew Dunstan                https://www.2ndQuadrant.com
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services




pgsql-hackers by date:

Previous
From: Bryn Llewellyn
Date:
Subject: Re: jsonb_object() seems to be buggy. jsonb_build_object() is good.
Next
From: Tomas Vondra
Date:
Subject: Re: 1 Status of vertical clustered index - 2 Join using(fk_constraint) suggestion - 3 Status of pgsql's parser autonomization