Robert Haas wrote
> On Mon, Dec 2, 2013 at 9:49 AM, Asit Mahato <
> rigid.asit@
> > wrote:
>> Hi all,
>>
>> I am a newbie. I am unable to understand the to do statement given below.
>>
>> Add full object name to the tag field. eg. for operators we need
>> '=(integer,
>> integer)', instead of just '='.
>>
>> please help me out with an example.
>>
>> Thanks and Regards,
>> Asit Mahato
>
> Cast the OID of the operator to regoperator instead of regoper.
This seems too simple an answer to be useful, and utterly confusing to the
OP. The ToDo item in question is in the pg_dump/pg_restore section. In
would seem to be possibly referring to this e-mail thread:
"Re: pg_dump sort order for functions"
http://www.postgresql.org/message-id/flat/9837222c1001120712t1e64eb2fg9f502757b600233e@mail.gmail.com#9837222c1001120712t1e64eb2fg9f502757b600233e@mail.gmail.com
though there is no link to prior discussion attached to the ToDo item. The
thread itself suggests there was yet prior discussion on the topic though my
quick search did not turn anything up.
It seems that not all objects (though it appears functions are currently one
exception) are fully descriptive in their tag/name output which make
deterministic ordering more difficult. The goal is, say for operators, to
output not only the base operator symbol (regoper) but the types associated
with the left-hand and right-hand sides (regoperator). The additional type
information makes the entire name unique (barring cross-schema conflicts at
least, which can be mitigated) and allows for deterministic ordering.
Asit, hopefully this gives you enough context to ask better questions which
others can answer more fully. Also, it may help to give more details about
yourself and your goals and not just throw out that you are a newbie. The
later gives people little guidance in how they should structure their help.
David J.
--
View this message in context:
http://postgresql.1045698.n5.nabble.com/Add-full-object-name-to-the-tag-field-tp5781167p5781431.html
Sent from the PostgreSQL - hackers mailing list archive at Nabble.com.