COGNOiSe.com - The IBM Cognos Community

IBM Cognos 10 Platform => Cognos 10 BI => Report Studio => Topic started by: dire on 31 Mar 2017 05:37:44 PM

Title: "Same Same Different" tests for potential duplicate payments
Post by: dire on 31 Mar 2017 05:37:44 PM
The purpose of the Same-Same-Different test (SSD) is to identify abnormal duplications as potential indicators of errors or fraud.

Any tips on how to apply this methodology in in Report Studio? I'd like to use [Invoice Number] and [Paid Amount] as the 'Same' factors, and [Vendor Number] as the 'Different' factor.

Consider the following data set:
Vendor Number   Paid Amount   Invoice Number
4   95.5   JAN002
4   95.5   JAN003
5   100.01   JAN004
4   100.01   FEB005
5   100.01   FEB005

I would like to return the following results:
Vendor Number   Paid Amount   Invoice Number
4   100.01   FEB005
5   100.01   FEB005


Any suggestions?

Thanks in advance!
Title: Re: "Same Same Different" tests for potential duplicate payments
Post by: dire on 05 Apr 2017 03:31:35 PM
Bump!!
Title: Re: "Same Same Different" tests for potential duplicate payments
Post by: dire on 06 Apr 2017 02:20:46 PM
Clarified question for ease - bump!
Title: Re: "Same Same Different" tests for potential duplicate payments
Post by: dire on 06 Apr 2017 03:51:08 PM
FYI, solved this problem doing the following:

1. Created a data item, [NEWDATAITEM], with the following expression: count(distinct [Vendor Number] for [Invoice Number], [Paid Amount])
2. Created a detail filter with the following expression: [NEWDATAITEM] > 1

This kicked out the instances where the invoice number and paid amounts were the same, but had distinct vendor numbers between the transactions.

Hopefully this helps someone in the future. :)