Comments on: IN vs INNER JOIN

By: IN vs INNER JOIN en Mysql. ¿Cuál es más rápida? | Grab This Code!

IN vs INNER JOIN en Mysql. ¿Cuál es más rápida? | Grab This Code! — Sun, 04 Mar 2012 21:07:49 +0000

[...] operadores en ese sistema gestor de base de datos. Sobre Sql Server existe una buena comparativa en esta página, la cual da como ganador a in versus inner join. Pero…y sobre Mysql? No he podido encontrar [...]

By: Gail

Gail — Mon, 22 Aug 2011 12:21:52 +0000

I’ve added the missing definition for BigTable.

By: Tom Thomson

Tom Thomson — Mon, 22 Aug 2011 11:40:16 +0000

I think that this statement is wreong:
—-quote
DECLARE @SomeTable (IntCol int)
Insert into @SomeTable (IntCol) Values (1)
Insert into @SomeTable (IntCol) Values (2)
Insert into @SomeTable (IntCol) Values (2)
Insert into @SomeTable (IntCol) Values (3)
Insert into @SomeTable (IntCol) Values (4)
Insert into @SomeTable (IntCol) Values (5)
Insert into @SomeTable (IntCol) Values (5)

SELECT *
FROM BigTable
WHERE SomeColumn IN (Select IntCol FROM @SomeTable)

This returns 5 rows and only columns from BigTable.
—-end quote
There’s no telling how many rows will be returned from BigTable – BigTable might have a 1000 rows where SomeColumn is 1, in which case the query will return at least 1000 rows; or it might have no rows with SomColumn less than 25, in which case it will return 0 rows. SOme people might be think you qactually meant what you wrote there, and be misled.

By: Nilesh Sane

Nilesh Sane — Fri, 06 May 2011 08:29:50 +0000

I agree, with the fact that if you populate the vaules in a table the subquery with IN will be slightly faster then an inner join. I was speaking from my experience where ‘a’ programmer, constructed the values which would form the part of the IN query by stringing together a comma seperated values (this method in itself is very inefficient). In such cases, if the programmer takes an approach of inserting such values in a table and doing an IN query, the performance will increase, as shown by you in this article.

By: Gail

Gail — Fri, 06 May 2011 08:09:57 +0000

In with value list is not really equivalent to an inner join, unless you take the time to populate the values into a table. If you do that, then the IN with subquery will likely still be faster (a little) than the inner join.

By: Nilesh Sane

Nilesh Sane — Fri, 06 May 2011 06:40:09 +0000

Thanks for sharing the info.
Theres another form of in query, where the in clause actually has the values which are comma separated. In such cases I have found that as the number of records in the in-clause increase the innner join query proves to be exponentially better than the in query. The difference becomes obvious when the number of records cross 100.

By: Gail

Gail — Wed, 04 May 2011 20:46:35 +0000

You mean a large number of matches, so each row in the one table has lots and lots of rows in the second?

In that case, you’d need to inner join and distinct to get rid of the duplicates that appear because of the full join. Adding a sort or hash to the execution plan is unlikely to make it faster.

By: Alex

Alex — Wed, 04 May 2011 18:45:41 +0000

Hello,

I simple explanation about your tests. IN was better than INNER JOIN ONLY because the result set you had used in your example had few registers. If you work with a large number of registers, and your table have unique indexs, INNER JOIN is probably better.
Try this out

By: Neha

Neha — Tue, 28 Sep 2010 06:48:15 +0000

Thanks ! Wonderful explanation. This helped resolve an issue!

By: dinesh

dinesh — Sun, 23 May 2010 15:57:17 +0000

Nice Article