From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
---|---|
To: | Michael Paquier <michael(at)paquier(dot)xyz> |
Cc: | Merlin Moncure <mmoncure(at)gmail(dot)com>, Dmitry Dolgov <9erthalion6(at)gmail(dot)com>, yarexeray(at)gmail(dot)com, pgsql-bugs(at)lists(dot)postgresql(dot)org |
Subject: | Re: BUG #15940: json_populate_recordset fails with ERROR: record type has not been registered |
Date: | 2019-08-19 22:04:25 |
Message-ID: | 30049.1566252265@sss.pgh.pa.us |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-bugs |
Michael Paquier <michael(at)paquier(dot)xyz> writes:
> On Tue, Aug 06, 2019 at 02:53:45PM -0500, Merlin Moncure wrote:
>> For posterity I agree that OP was essentially exploiting undefined, or
>> at least poorly defined, behavior. For my money, I'd advise using
>> this function for cases where you don't want to use an in place type,
>> just a column list:
> If I were to change something here, that would be this error string
> exposed to the user. With the current error, there is no real way
> that the user knows what is wrong and why he/she should not do that.
> Could it be possible to add some recommendation for example on top of
> an error "cannot do that because blah"?
I concluded that we'd better restore the former behavior, or people
will complain that this is a regression. We can however do better
than the "record type has not been registered" error, at least for
the case where the error is being thrown at runtime. I went with
ERROR: could not determine row type for result of json_populate_record
HINT: Provide a non-null record argument, or call the function in the FROM clause using a column definition list.
regards, tom lane
From | Date | Subject | |
---|---|---|---|
Next Message | PG Bug reporting form | 2019-08-20 06:52:12 | BUG #15967: Sequence generation using NEXTVAL() fails on 64bit systems |
Previous Message | Bansal, Abhishek (Abhishek) | 2019-08-19 20:30:54 | RE: Postgresql Query |