Hi folks,
While performing Inner Join, the rows containing the null field in the join
predicate are dropped. Sometimes, it is required that the rows containing
join predicate as null are not dropped from the output, while performing
inner join.
Is there any way how this can be achieved?
A possible workaround: The join keys which are null can be replaced with
any non-null value, preferably a chararray 'NULL'. The inner join would
then retain this row.
Is it a better option to introduce an operator Identical to existing
JOIN(Inner), which would also handle nulls in an inner join?
NULLJOIN will take care of marshalling and unmarshalling the null values
with non-null value and converting this back to null, after the join
operation is complete.
thanks
deepak marathe
While performing Inner Join, the rows containing the null field in the join
predicate are dropped. Sometimes, it is required that the rows containing
join predicate as null are not dropped from the output, while performing
inner join.
Is there any way how this can be achieved?
A possible workaround: The join keys which are null can be replaced with
any non-null value, preferably a chararray 'NULL'. The inner join would
then retain this row.
Is it a better option to introduce an operator Identical to existing
JOIN(Inner), which would also handle nulls in an inner join?
NULLJOIN will take care of marshalling and unmarshalling the null values
with non-null value and converting this back to null, after the join
operation is complete.
thanks
deepak marathe