In this chapter we move to the most important component of the ETL design in Data Services: the dataflow object. The dataflow object is the container that holds all transformations that can be performed on data.
The structure of the dataflow object is simple: one or many source objects are placed, on the left-hand side (which we extract the data from), then source objects are linked to the series of transform objects (which perform manipulation on the data extracted), and finally, the transform objects are linked to one or many target table objects (telling Data Services where the transformed data should be inserted). During the transformation of the dataset inside the dataflow, you can split the dataset into multiple dataset flows, or conversely, merge multiple separately transformed dataflows together.
Manipulations performed on data inside dataflows are done on a row-by-row basis. The rows extracted from the source go from left to right through all objects placed inside the...