From: Roy Lyseng
Date: August 25 2010 6:43pm
Subject: Re: bzr commit into mysql-5.1-bugteam branch (martin.hansson:3478)
 Bug#51070
List-Archive: http://lists.mysql.com/commits/116796
Message-Id: <4C756441.7080404@oracle.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit


>> +SELECT * FROM t1 WHERE ( a, b ) IN ( SELECT c, d FROM t2 ) IS NULL;
>> +SELECT * FROM t1 WHERE ( a, b ) NOT IN ( SELECT c, d FROM t2 ) IS UNKNOWN;
>> +SELECT * FROM t1 WHERE (( a, b ) NOT IN ( SELECT c, d FROM t2 )) IS UNKNOWN;
>
> as you add (), is it that
> ( a, b ) NOT IN ( SELECT c, d FROM t2 ) IS UNKNOWN
> stands for
> ( a, b ) NOT IN (( SELECT c, d FROM t2 ) IS UNKNOWN)
> ? Mmmm no that wouldn't be parsable, (c,d) can't be compared to UNKNOWN.
> Then maybe the added () are "just for fun"?

They are different, but equivalent, syntactic constructs, so it can have some 
value in checking that the parser does a proper job.

They are both incarnations of:
<boolean test> ::= <boolean primary> [ IS [ NOT ] <truth value> ]

But in the first case, the <boolean primary> is a <predicate>, more precisely an
<in predicate>.

In the second case, the <boolean primary> is a <boolean predicand>, which 
traslates to a <parenthesized boolean value expression>, which is
<parenthesized boolean value expression>, which again is
<left paren> <boolean value expression> <right paren>.

Finally, the <boolean value expression> is translated to a <predicate>, as above.

 From your SQL oracle...