Peter Eisentraut wrote:
> Bruce Momjian writes:
> > But it isn't a message:
> >
> > Table "public.xx"
> > Column | Type | Modifiers
> > --------+---------+-----------
> > y | integer |
> > Indexes:
> > "ii" btree (y)
>
> By definition, everything that is printed out for human consumption is a
> message.
>
> Everything that is printed out for consumption by an SQL parser (e.g., in
> pg_dump) is encouraged to use SQL quoting rules. But everything that is
I see pg_dump using the same rules that I am proposing:
CREATE TABLE "xx y" (
y integer
) WITH OIDS;
CREATE TABLE xx (
y integer
) WITH OIDS;
[ We need that WITH OIDS fixed before 7.5. I think Neil is working on it.]
> intended to be read by humans should use quoting rules that are common in
> real-life publishing.
To me quoting the table name is like quoting the name of a section
heading (which people don't do), and that's really what we are
displaying:
test=> \d xx
Table public.xx
Column | Type | Modifiers
--------+---------+-----------
y | integer |
Indexes:
vv btree (y)
"Table public.xx" is the section heading, and before you complain that I
quoted it in this sentence, I had to because it is part of a sentence,
not on its own as it is in psql \d display.
You were suggesting no quotes at all. Would this display be OK?
test=> \d "xx y"
Table public.xx y
Column | Type | Modifiers
--------+---------+-----------
y | integer |
Indexes:
vv btree (y)
Note I had to quote the table name to get \d to accept it. To me that
'y' just sitting along looks strange.
--
Bruce Momjian | http://candle.pha.pa.us
pgman@candle.pha.pa.us | (610) 359-1001
+ If your life is a hard drive, | 13 Roberts Road
+ Christ can be your backup. | Newtown Square, Pennsylvania 19073
