Worked Examples - Getting Started with Talend Open Studio for Data Integration - page 267

We also need to extract the delivery SKUs, but cannot simply send all of the delivery
SKUs to Fabulous Fashions, because not all orders will contain Fabulous Fashion items
and so, not all delivery SKUs will be relevant to Fabulous Fashions. In the second flow
of the subjob, we extract all of the order lines and filter them so that only the delivery
lines are passed through tFilterRow_3 as shown in the following screenshot:
We then join this dataflow with a lookup dataflow which only contains order lines
for Fabulous Fashions SKUs. The lookup flow is, in fact, an exact copy of the first
flow described earlier. In the tMap component, we join the two flows on Order ID,
so the data flow emerging from the tMap component is all of the delivery SKUs that
appear in orders that also contained at least one FAB SKU.
Why do we have the same flow duplicated in a job? In this case, the
first flow of FAB order lines and the lookup flow of the same. The
reason is that the Studio won't allow you to replicate a dataflow to
different points in another dataflow, which is what would be required
if we were to achieve this with a single dataflow of FAB order lines.
In order to get the desired result, we need to undertake two separate
and identical reads of the input file.