If you use the Lookup Transform to perform lookups with columns that can be NULL, the Lookup Transform will fail to find expected matches because in a database world, NULL does not equal NULL.
So how do we make the Lookup Transform match as expected when source column and lookup column can be NULL? We use ISNULL function, partial caching and a custom query in the Lookup Transform.
Let us look at a quick example. This is our lookup table definition, which allows NULLs in the LastName column, that we want to get the Id of a person from, based on their first and last name:
CREATE TABLE Person ( Id int IDENTITY(1,1) NOT NULL, FirstName varchar(50) NOT NULL, LastName varchar(50) NULL )
In SSIS, we have a input source with a
FirstName column and a
LastName, where the
LastName column might sometimes be NULL.
To find the
Id of each name, we add a Lookup Transform to the Data Flow Task, and configure it to use Partial cache in the General tab. In the Connection tab we point to our
Person table as normal, and in the Columns tab we map our input’s
FirstName column and
LastName column to the lookup’s
Until now, every step has been the same as with Full caching. However, now we continue to the Advanced tab and check the *Modify the SQL statement checkbox and change the SQL like so:
select * from (select * from [Person]) [refTable] where [refTable].[FirstName] = ? and ISNULL([refTable].[LastName], '') = ISNULL(?, '')
With the help of the
ISNULL function, we rewrite any NULL values to an empty string, which is comparable, but only if the LastName is actually NULL.
Last step is to click the Parameters button and assign the parameters.
Hope this help!