COGNOiSe.com - The IBM Cognos Community

ETL, App Development and other tools => COGNOS DecisionStream/Data Manager => Topic started by: gplsupport on 10 Aug 2006 05:22:09 PM

Title: Handling of data exceptions during Decision Stream transformations
Post by: gplsupport on 10 Aug 2006 05:22:09 PM
Hi all,

We are loading files supplied by external parties using DS.  The externally supplied files sometimes contain errors  (eg date expected as DD/MM/YYYY but supplied as YYYY/MM/DD).  This causes the DS load process to fail.

Is there a way of DS identifying the record and field that it failed on without having to handraulically review the input file?

Cheers
...David S
Title: Re: Handling of data exceptions during Decision Stream transformations
Post by: MFGF on 29 Aug 2006 06:31:23 AM
Instead of handling these as date datatypes on input, bring them in as char strings, and validate them against a lookup of valid dates in the same format (ie char datatype, DD/MM/YYYY format).  Your lookup could also contain a second attribute with the same dates defined as date datatypes.  Any rows coming into your build with invalid dates will then be rejected, but the good ones will make it through.  You can add a transformation model derivation to get hold of the true date from the second attribute in the lookup, and pass this through to your delivery table(s).

Just a thought...

MF.
Title: Re: Handling of data exceptions during Decision Stream transformations
Post by: gplsupport on 29 Aug 2006 02:11:24 PM
Thanks for your reply!  I tried using the documented options in SQLTXT but they didn't work (confirmed with Cognos support).  I will implement your suggestion.