Re: "AS" by the syntax of table reference.(8.4 proposal)

"Hiroshi Saito" <z-saito(at)guitar(dot)ocn(dot)ne(dot)jp> writes:
> *** src/backend/parser/gram.y 4 Feb 2008 06:35:46 -0000 1.1
> --- src/backend/parser/gram.y 4 Feb 2008 17:24:19 -0000 1.2
> ***************
> *** 8320,8325 ****
> --- 8320,8333 ----
> $$->val = (Node *) $1;
> $$->location = @1;
> }
> + | c_expr IDENT
> + {
> + $$ = makeNode(ResTarget);
> + $$->name = $2;
> + $$->indirection = NULL;
> + $$->val = (Node *) $1;
> + $$->location = @1;
> + }
> | a_expr
> {
> $$ = makeNode(ResTarget);

Hmm, since c_expr is so restrictive, is that really going to satisfy
anybody who expects to be able to omit AS? With the number of
unreserved keywords we have, the restriction to IDENT for the label
seems like it's going to be a constant source of gotchas as well. Both
of these limitations are contrary to the SQL spec, too --- SQL99 says
that a SELECT-list element is

<derived column> ::=
<value expression> [ <as clause> ]

<as clause> ::= [ AS ] <column name>

so we really need to support at least ColId as the allowed set of
column alias names. (I tried changing the patch to do that, but
got a lot of shift/reduce conflicts, some of which are maybe fixable
but some seem hard to fix.)

I think that simply saying "it's not supported" is going to be easier to
deal with than people trying to use it and complaining that this or that
case doesn't work.

regards, tom lane