Re: postgres question: Views with duplicate field names

On 09/05/2016 12:55 PM, Ryan Murphy wrote:
> Hello, I have a question about views in Postgres.
>
> Given a table like so:
>
> create table todo (
> id serial,
> task text,
> done_time timestamp default null
> );
>
> it is legal (though perhaps not advised, by some) to query it like so:
>
> select task, * from todo;
>
> This gives a result with 2 redundant "task" fields (with duplicate names):
>
> task | id | task | done_time
> --------------+----+--------------+-----------
> wash the dog | 1 | wash the dog |
>
> However, if I try to make a view of this I hit a problem: views can't
> have duplicate field names:
>
> create view task2 as select task, * from todo;
>
> ERROR: column "task" specified more than once
>
> I understand this may seem like a silly thing to want to do, but my
> question is if there is an easy way to automatically de-dup the columns
> of the query so I can create a view from it. Or is there any
> fundamental reason why views can't be allowed to have duplicate columns,
> just like the result set above?

test=> create view task2 as select task AS task_1 , * from todo;
CREATE VIEW

test=> \d task2
View "public.task2"

Column | Type | Modifiers

-----------+-----------------------------+-----------

task_1 | text |

id | integer |
task | text |
done_time | timestamp without time zone |

>
> Thanks!
>
> Ryan

--
Adrian Klaver
adrian(dot)klaver(at)aklaver(dot)com