From: | Raphael Araújo e Silva <raphael(at)pgmodeler(dot)com(dot)br> |
---|---|
To: | pgsql-sql(at)postgresql(dot)org |
Subject: | Confusion about CREATE OPERATOR syntax. |
Date: | 2017-09-05 18:38:03 |
Message-ID: | CABJ1nXxMdZWEsSUphsfrT-vOZZxRo3ZMeaOA9WS=5KHEvv5fzg@mail.gmail.com |
Views: | Whole Thread | Raw Message | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-sql |
Hi guys!
I'm the creator of the PostgreSQL Database Modeler and I'm improving the
reverse engineering feature after some requests from the users.
I've stumbled on a bug when importing operators and while I was writing a
fix for that bug a great confusion came into my mind. It's about the CREATE
OPERATOR syntax, more specifically, about the LEFTARG and RIGHTARG
attributes.
I've read the docs but cannot find a proper answer and that's why I'm
here... I hope you guys can help me. Well, my doubt is: when specifying
LEFTARG and RIGHTARG the datatypes should be written in the complete form?
By "complete form" I mean specifying the precision and/or length... For
instance, I want to create an operator to handle certain operation between
two numeric(5,2) values, said that, my operator should be written as:
CREATE OPERATOR ++ (
PROCEDURE = func_name(numeric,numeric),
LEFTARG = numeric,
RIGHTARG = numeric
)
Or(?):
CREATE OPERATOR ++ (
PROCEDURE = func_name(numeric,numeric),
LEFTARG = *numeric(5,2)*,
RIGHTARG = *numeric(5,2)*
)
Well, I've tested both forms and them work fine. If I try to run one after
another an error will raise for the second reporting that the operator
already exists. Is PostgreSQL simply interpreting them as the same? Is the
precision/length ignored/not needed?
Thanks in advance and forgive my rusty English, I'm Brazilian! :)
--
Raphael Araújo e Silva
pgModeler Project Developer
raphael(at)pgmodeler(dot)com(dot)br
http://pgmodeler.com.br
https://github.com/pgmodeler
From | Date | Subject | |
---|---|---|---|
Next Message | David G. Johnston | 2017-09-06 00:31:08 | Re: Confusion about CREATE OPERATOR syntax. |
Previous Message | Viktor Bojović | 2017-09-01 07:59:16 | Re: autogenerated json path |