Data type conversions RRS feed

  • Question

  • What are the pros/cons of datatype conversions using the advanced properties of the source adapters or transformation components themselves vs. using the data conversion transformation?  Sometimes I am able to do all the conversions I need using the advanced properties rather than use a data conversion component, but didn't know if this was considered bad practice, or has some other impact that isn't immediately apparent.



    Monday, September 25, 2006 3:07 AM

All replies

  • I offten use Derived Column component as data conversion

    I think that is well used to convert lots of data type at the same time

    Monday, September 25, 2006 4:06 AM
  • I'd say doing as much conversion up front in a source as possible would be preferable. Doing it later on means you have a wider buffer, to accomodate the two columns required. Saying that, they can be harder to maintain sometimes, as refreshing the metadata can loose your little tweaks. Depends on how stable you view your source perhaps.

    Using one transformation over another, Derived Column or Data Conversion would not seem like much of a choice really, I doubt there is much to choose between. They are both synchronous transforms, so will require two "columns" in the buffer, the current type and the new type. That obviously makes for a wider buffer and therefore more cost, due to less rows per buffer.

    Still do what works for you, but test a few scenarios if you can and see which works best for you. If performance or maintanece a bigger cost to you?

    Monday, September 25, 2006 6:49 AM