A message was posted on the SSIS forum earlier today asking about ExclusionGroups. BOL is a little light on the subject of why you might use Exclusion Groups so I thought I would elaborate here a little.
An ExclusionGroup can be used by synchronous components that have more than one output. In short, an ExclusionGroup is used to tell the pipeline engine that each incoming row will only go to ONLY ONE of the outputs. If you don't use a non-zero ExclusionGroup then the row will go to ALL outputs (in much the same way as happens with the Multitask component) and that probably isn't what you want.
The forum poster said the following:
I have a question about this because I have never used the "ExclusionGroup" property. For example, I have a script component where I specify 4 separate outputs, because I am sending different groups of rows to each output. I accomplish this programmatically using a lot of conditionals and it works fine.
I did not have to use the "ExclusionGroup" property to do this. So I'm not sure why I would ever need this, or to use DirectRow? I'm trying to understand this better, because maybe I feel like I'm not understanding the DirectRow, or how/when to use it.
I can only presume that her 4 outputs were not synchronous to the input (i.e. they are asynchronous) and hence an ExclusionGroup was not needed, in fact it would be irrelevant. Synchronous components will generally perform quicker than an asynchronous component that does the same thing hence the use of an ExclusionGroup can be very valuable. In September 2005 I posted a blog entry demonstrating the use of an ExclusionGroup. You can read it here:
Multiple outputs from a synchronous script transform
(http://blogs.conchango.com/jamiethomson/archive/2005/09/05/SSIS-Nugget_3A00_-Multiple-outputs-from-a-synchronous-script-transform.aspx)
Any more questions about Exclusion Groups? Just let me know or, better still, post your question on the SSIS forum.
-Jamie