Re: Using row_to_json with %ROWTYPE ?

> So either PostgreSQL is seeing a different view (in a different schema) or the function is confused in ways difficult to predict.

Seriously ? You still want to continue calling it user-error ? There
is no other view, there is no other schema , I am not hiding anything
from you !

On 5 February 2015 at 23:38, David Johnston <david(dot)g(dot)johnston(at)gmail(dot)com> wrote:
> On Thu, Feb 5, 2015 at 4:29 PM, Tim Smith <randomdev4+postgres(at)gmail(dot)com>
> wrote:
>>
>> You're most welcome to look at my view definition view if you don't
>> believe me ....
>>
>> View definition:
>> SELECT a.session_id,
>> a.session_ip,
>> a.session_user_agent,
>> a.session_start,
>> a.session_lastactive,
>> b.user_id,
>> b.tenant_id,
>> b.reseller_id,
>> b.tenant_name,
>> b.user_fname,
>> b.user_lname,
>> b.user_email,
>> b.user_phone,
>> b.user_seed,
>> b.user_passwd,
>> b.user_lastupdate,
>> b.tenant_lastupdate
>> FROM app_sessions a,
>> app_users_vw b
>> WHERE a.user_id = b.user_id;
>
>
> So that view and definition are correct.
>
> So either PostgreSQL is seeing a different view (in a different schema) or
> the function is confused in ways difficult to predict.
>
> I guess it is possible that:
>
> (SELECT v_row FROM v_row) would give that message but I get a "relation
> v_row does not exist" error when trying to replicate the scenario.
>
> It may even be a bug but since you have not provided a self-contained test
> case, nor the version of PostgreSQL, the assumption is user error.
>
> David J.
>