Thread: Difference between IN and JOIN

Hi,

I lately wondered if there is a difference between a JOIN and a IN in
queries similar to the following:

SELECT f1 FROM t1 JOIN t2 ON (t.f2 = t2.f2) WHERE t2.f3 = x

SELECT f1 FROM t1 WHERE t1.f2 IN (SELECT f2 FROM t2 WHERE f3 = x)

As I see it there's no semantic difference between the two. Are there
differences in performance or anything else that matters? The execution
plans seem to match except for the use of an "Hash IN Join" in place of
a "Hash Join". Estimated costs match.

Thanks for your advice

-- 
Markus Bertheau <twanger@bluetwanger.de>

Markus Bertheau <twanger@bluetwanger.de> writes:
> I lately wondered if there is a difference between a JOIN and a IN in
> queries similar to the following:

> SELECT f1 FROM t1 JOIN t2 ON (t.f2 = t2.f2) WHERE t2.f3 = x

> SELECT f1 FROM t1 WHERE t1.f2 IN (SELECT f2 FROM t2 WHERE f3 = x)

> As I see it there's no semantic difference between the two.

There's plenty of difference, if t2 contains repeated occurrences of f2
values.
        regards, tom lane

В Пнд, 04.10.2004, в 16:17, Tom Lane пишет:
> Markus Bertheau <twanger@bluetwanger.de> writes:
> > I lately wondered if there is a difference between a JOIN and a IN in
> > queries similar to the following:
>
> > SELECT f1 FROM t1 JOIN t2 ON (t.f2 = t2.f2) WHERE t2.f3 = x
>
> > SELECT f1 FROM t1 WHERE t1.f2 IN (SELECT f2 FROM t2 WHERE f3 = x)
>
> > As I see it there's no semantic difference between the two.
>
> There's plenty of difference, if t2 contains repeated occurrences of f2
> values.

In my case f2 is the primary key of t2. I probably didn't state the
point of my question well enough: You can write an IN query as a
semantically equal JOIN. I wondered, which one I should prefer, and on
which grounds.

Thanks

--
Markus Bertheau <twanger@bluetwanger.de>