From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
---|---|
To: | Dennis Bjorklund <db(at)zigo(dot)dhs(dot)org> |
Cc: | pgsql-hackers(at)postgresql(dot)org |
Subject: | Re: IN and ANY |
Date: | 2004-03-02 06:45:02 |
Message-ID: | 24070.1078209902@sss.pgh.pa.us |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
Dennis Bjorklund <db(at)zigo(dot)dhs(dot)org> writes:
> Hmm, the draft seems to be broken since I can only find ANY defined for
> subqueries in other sections, and not for value lists. Strange but not
> uncommon. Now I don't know what the standard says about this. Maybe
> someone with the sql99 spec wants to check.
I think you are reading the term "equivalent" as meaning an equivalence
in both directions. It looks to me that the spec's definition of
<in predicate> is (mis)using the term to mean "is defined as".
In SQL92 I see
1) Let IVL be an <in value list>.
( IVL )
is equivalent to the <table value constructor>:
( VALUES IVL )
...
4) The expression
RVC IN IPV
is equivalent to
RVC = ANY IPV
These two rules together define both forms of IN in terms of the
"= ANY (subquery)" construct. But surely the first rule is not
meant to say that VALUES is a noise word. So this has to be a
one-way implication.
Accordingly I think you are in error to suggest that "= ANY (valuelist)"
is supposed to work. I think ANY is only supposed to have a table
subquery to the right.
I don't have a strong opinion about "IN array", but am worried that
allowing it would create ambiguity about which interpretation is meant.
Is the left-hand side supposed to be compared against the whole array or
each array member?
regards, tom lane
From | Date | Subject | |
---|---|---|---|
Next Message | Shridhar Daithankar | 2004-03-02 06:58:22 | Re: Avoid MVCC using exclusive lock possible? |
Previous Message | Alvaro Herrera | 2004-03-02 06:21:32 | Re: Pl/Java - next step? |