The Data Integration Service applies the semi-join optimization method to a Joiner transformation when one input group has many more rows than the other and when the larger group has many rows with no match in the smaller group based on the join condition. The Data Integration Service attempts to decrease the size of the data set of one join operand by reading the rows from the smaller group, finding the matching rows in the larger group, and then performing the join operation. Decreasing the size of the data set improves mapping performance because the Data Integration Service no longer reads unnecessary rows from the larger group source. The Data Integration Service moves the join condition to the larger group source and reads only the rows that match the smaller group.

Before applying the semi-join optimization method, the Data Integration Service performs analyses to determine whether semi-join optimization is possible and likely to be worthwhile. If the analyses determine that this method is likely to improve performance, the Data Integration Service applies it to the mapping. The Data Integration Service then reanalyzes the mapping to determine whether there are additional opportunities for semi-join optimization. It performs additional optimizations if appropriate.

The Developer tool does not enable this method by default.