Greg Stark <gsstark(at)mit(dot)edu> writes:
> Uhm, doesn't the spec have anything to say about this?
> I mean, the view sure looks like standard SQL on its face.
Well, you might read the spec as requiring the view column to have
datatype CHAR(n) where n is the length of the unknown literal.
I see in SQL92:
9) The data type of a <character string literal> is fixed-length
character string. The length of a <character string literal>
is the number of <character representation>s that it contains.
Note: <character string literal>s are allowed to be zero-length
strings (i.e., to contain no characters) even though it is
not permitted to declare a <data type> that is CHARACTER with
The NOTE would appear to imply that
CREATE VIEW v AS SELECT '';
is illegal, which is not a conclusion I care to follow blindly. In any
case, in Postgres I would think we would want to take the type as "text"
not "CHAR(n)", spec or no spec.
regards, tom lane
In response to
pgsql-hackers by date
|Next:||From: Andrew Dunstan||Date: 2003-10-03 18:32:56|
|Previous:||From: Gottfried F. Zojer||Date: 2003-10-03 17:22:39|
|Subject: Re: Question regarding coopting Database Engine|