Hi,

in my dataflow, I've 5 columns in the input of a 'Derived column' task.
3 of these 5 columns allow me to compute the value of a new column. Then, I
don't need anymore these columns in my pipeline.

How is it possible to keep only the columns I need ?

Maybe I'm blind, but I can't find a 'Select column' task ...

I've got the same problem after a conversion task, I don't want to keep the
value of the old type.


Thanks !


Oxiane
----------------------------------------------------------
Omnia vanitas

You cannot, nor do you actually need to do anything. The cost of changing the
buffer size probably outweighs the cost carrying them through, well that is
my impression. Certain transforms, normally asynchronous in nature, create
new buffers and at that point you can be more selective. Examples include
Aggregate, Sort.

Darren

--
http://wiki.sqlis.com | http://www.sqlis.com | http://www.sqldts.com


"Timur" wrote:

Timur,
Think of it this way. Those values are in the buffer anyway at runtime so
its actually more expensive to explicitly get rid of them.

I can see value in "hiding" columns from downstream components mind you
although do bear in mind that this would only make the GUI more readable, it
wouldn't alter the performance of the data-flow. If you think this feature
would be useful then vote for it at the product feedback centre:
http://lab.msdn.microsoft.com/produc...k/default.aspx

-Jamie


"Timur" wrote: