Re: Re: UNION JOIN vs UNION SELECT

Tom Lane <tgl@sss.pgh.pa.us>

From: Tom Lane <tgl@sss.pgh.pa.us>
To: chris@bitmead.com
Cc: pgsql-hackers@postgreSQL.org
Date: 2000-08-29T00:27:27Z
Lists: pgsql-hackers
Chris <chrisb@nimrod.itg.telstra.com.au> writes:
>> the grammar is just plain not LR(1) unless you
>> count UNION JOIN as a single token. 

> Would it be bad to make UNION JOIN as a single token?

That's exactly the solution I'm proposing.  However, it's pretty painful
to make the lexer do it directly (consider intervening comments, for
example) so what I have in mind is a filter between the parser and lexer
that does one-token lookahead when it finds a UNION token.  If next
token is JOIN, pass back just one UNIONJOIN token, else stash away the
second token to be returned on next call from parser.

			regards, tom lane