NOT IN vs NOT EXISTS
Which of these queries is the faster?
SELECT ProductID, ProductName FROM Northwind..Products p WHERE NOT EXISTS ( SELECT 1 FROM Northwind..[Order Details] od WHERE p.ProductId = od.ProductId)
Or NOT IN:
SELECT ProductID, ProductName FROM Northwind..Products p WHERE p.ProductID NOT IN ( SELECT ProductID FROM Northwind..[Order Details])
The query execution plan says they both do the same thing. If that is the case, which is the recommended form?
This is based on the NorthWind database.
Just found this helpful article: http://weblogs.sqlteam.com/mladenp/archive/2007/05/18/60210.aspx
I think I'll stick with NOT EXISTS.
I always default to
The execution plans may be the same at the moment but if either column is altered in the future to allow
NOT IN version will need to do more work (even if no
NULLs are actually present in the data) and the semantics of
NOT IN if
NULLs are present are unlikely to be the ones you want anyway.
[Order Details].ProductID allow
NOT IN will be treated identically to the following query.
SELECT ProductID, ProductName FROM Products p WHERE NOT EXISTS (SELECT * FROM [Order Details] od WHERE p.ProductId = od.ProductId)
The exact plan may vary but for my example data I get the following.
A reasonably common misconception seems to be that correlated sub queries are always "bad" compared to joins. They certainly can be when they force a nested loops plan (sub query evaluated row by row) but this plan includes an anti semi join logical operator. Anti semi joins are not restricted to nested loops but can use hash or merge (as in this example) joins too.
/*Not valid syntax but better reflects the plan*/ SELECT p.ProductID, p.ProductName FROM Products p LEFT ANTI SEMI JOIN [Order Details] od ON p.ProductId = od.ProductId
[Order Details].ProductID is
NULL-able the query then becomes
SELECT ProductID, ProductName FROM Products p WHERE NOT EXISTS (SELECT * FROM [Order Details] od WHERE p.ProductId = od.ProductId) AND NOT EXISTS (SELECT * FROM [Order Details] WHERE ProductId IS NULL)
The reason for this is that the correct semantics if
[Order Details] contains any
ProductIds is to return no results. See the extra anti semi join and row count spool to verify this that is added to the plan.
Products.ProductID is also changed to become
NULL-able the query then becomes
SELECT ProductID, ProductName FROM Products p WHERE NOT EXISTS (SELECT * FROM [Order Details] od WHERE p.ProductId = od.ProductId) AND NOT EXISTS (SELECT * FROM [Order Details] WHERE ProductId IS NULL) AND NOT EXISTS (SELECT * FROM (SELECT TOP 1 * FROM [Order Details]) S WHERE p.ProductID IS NULL)
The reason for that one is because a
Products.ProductId should not be returned in the results except if the
NOT IN sub query were to return no results at all (i.e. the
[Order Details] table is empty). In which case it should. In the plan for my sample data this is implemented by adding another anti semi join as below.
The effect of this is shown in the blog post already linked by Buckley. In the example there the number of logical reads increase from around 400 to 500,000.
Additionally the fact that a single
NULL can reduce the row count to zero makes cardinality estimation very difficult. If SQL Server assumes that this will happen but in fact there were no
NULL rows in the data the rest of the execution plan may be catastrophically worse, if this is just part of a larger query, with inappropriate nested loops causing repeated execution of an expensive sub tree for example.
This is not the only possible execution plan for a
NOT IN on a
NULL-able column however. This article shows another one for a query against the
NOT IN on a
NOT NULL column or the
NOT EXISTS against either a nullable or non nullable column it gives the following plan.
When the column changes to
NOT IN plan now looks like
It adds an extra inner join operator to the plan. This apparatus is explained here. It is all there to convert the previous single correlated index seek on
Sales.SalesOrderDetail.ProductID = <correlated_product_id> to two seeks per outer row. The additional one is on
WHERE Sales.SalesOrderDetail.ProductID IS NULL.
As this is under an anti semi join if that one returns any rows the second seek will not occur. However if
Sales.SalesOrderDetail does not contain any
ProductIDs it will double the number of seek operations required.
Read more... Read less...
Also be aware that NOT IN is not equivalent to NOT EXISTS when it comes to null.
This post explains it very well
When the subquery returns even one null, NOT IN will not match any rows.
The reason for this can be found by looking at the details of what the NOT IN operation actually means.
Let’s say, for illustration purposes that there are 4 rows in the table called t, there’s a column called ID with values 1..4
WHERE SomeValue NOT IN (SELECT AVal FROM t)
is equivalent to
WHERE SomeValue != (SELECT AVal FROM t WHERE ID=1) AND SomeValue != (SELECT AVal FROM t WHERE ID=2) AND SomeValue != (SELECT AVal FROM t WHERE ID=3) AND SomeValue != (SELECT AVal FROM t WHERE ID=4)
Let’s further say that AVal is NULL where ID = 4. Hence that != comparison returns UNKNOWN. The logical truth table for AND states that UNKNOWN and TRUE is UNKNOWN, UNKNOWN and FALSE is FALSE. There is no value that can be AND’d with UNKNOWN to produce the result TRUE
Hence, if any row of that subquery returns NULL, the entire NOT IN operator will evaluate to either FALSE or NULL and no records will be returned
If the execution planner says they're the same, they're the same. Use whichever one will make your intention more obvious -- in this case, the second.
Actually, I believe this would be the fastest:
SELECT ProductID, ProductName FROM Northwind..Products p outer join Northwind..[Order Details] od on p.ProductId = od.ProductId) WHERE od.ProductId is null
I have a table which has about 120,000 records and need to select only those which does not exist (matched with a varchar column) in four other tables with number of rows approx 1500, 4000, 40000, 200. All the involved tables have unique index on the concerned
NOT IN took about 10 mins,
NOT EXISTS took 4 secs.
I have a recursive query which might had some untuned section which might have contributed to the 10 mins, but the other option taking 4 secs explains, atleast to me that
NOT EXISTS is far better or at least that
EXISTS are not exactly the same and always worth a check before going ahead with code.
In your specific example they are the same, because the optimizer has figured out what you are trying to do is the same in both examples. But it is possible that in non-trivial examples the optimizer may not do this, and in that case there are reasons to prefer one to other on occasion.
NOT IN should be preferred if you are testing multiple rows in your outer select. The subquery inside the
NOT IN statement can be evaluated at the beginning of the execution, and the temporary table can be checked against each value in the outer select, rather than re-running the subselect every time as would be required with the
NOT EXISTS statement.
If the subquery must be correlated with the outer select, then
NOT EXISTS may be preferable, since the optimizer may discover a simplification that prevents the creation of any temporary tables to perform the same function.