December 28, 2009 at 12:50 am
Hi,
I'm working on an ETL package (General Ledger) where I use a Conditional Split to split the input into two outputs (O1 and O2) based on the Account Type. Below the Conditional Split, I use Derived Column transformations for both O1 and 02. The Derived Column transformation looks like this:
O1:
ETLDT: @[User::CETDT]
GroupNum: 11
Flag: 1
GLAmount: (DebitAmount - CreditAmount)
O2:
ETLDT: @[User::CETDT]
GroupNum: 11
Flag: 1
GLAmount: (CreditAmount - DebitAmount)
As you can see, the Derived Columns ETLDT, GroupNum and Flag are common in both O1 and O2. Also, they do not impact the Conditional Split in any way.
I am a little confused about whether I should use the same approach as mentioned above, or if I should add a separate Derived Column above the Conditional Split for the ETLDT, GroupNum and Flag derived columns, <b>in order to get optimal performance</b>.
Kindly help!
December 28, 2009 at 1:53 am
What is criteria for arriving Debit Amount and Credit Amount?
December 28, 2009 at 2:31 am
I believe this approach is fine because it may not make sense manually adding the condition for account type (which would be required if derived column transform done above conditional split transform) when the conditional split transform is already doing that. Secondly the number of operations are going to be the same and are not duplicated for both the conditions.
December 31, 2009 at 1:15 am
Thanks for the response, guys! I eventually discarded this approach and went ahead with 'Debit - Credit' regardless of Account Type. This, I was told by a few Finance guys, was the way to go from a functional perspective.
But I guess this post would be useful in the future for those who face similar situations when working on ETL for other industries.
Thanks a ton, guys! I really appreciate it. 🙂
Viewing 4 posts - 1 through 3 (of 3 total)
You must be logged in to reply to this topic. Login to reply