The most important task in SSIS is the Data Flow Task. The SSIS Data Flow Task can be selected
directly from the SSDT Toolbox, and then the source and destinations are defined within the task.
The Data Flow Task isn’t merely a mapping transform for input and output columns. This task has
its own design surface like the Control Flow, where you can arrange task-like components called
transforms to manipulate data as it flows in a pipeline from the source to a destination. The Data
Flow, as you can imagine, is the heart of SSIS, because it encapsulates all the data transformation
aspects of ETL.
Data Flows can split the data in the pipeline based on a data element and handle each stream
separately. In the Data Flow, the header line of the file can be split off and examined separately from
the detail lines. As the pipeline exits the data-cleansing process, the streams can be sent to separate
destinations or converged to a final combined destination. Note that you may have several different
Data Flows within an SSIS package. For each of the Data Flow Tasks you add to the control surface,
you’ll have a corresponding Data Flow surface. This task is so important and such an integral part
of moving data in SSIS that it is covered in detail in Chapter 4.