Trying to find the lowest value per Dimension, possibly aggr()?

[Anonymous]

Hello,

I have a chart and I need to identify the lower Incentive value for each %ADAcctKey.

The Incentive value is a calculated value.  In the example below 445 would flag 10% and for 2544 the lowest flagged would be 0%.

%ADAcctKey	MetricDesc	QA	Adherence	Revenue	Saves	Level	Target	Incentive
445	QA	87.50%				Target	80% >=	10%
445	Adherence		88.04%			Target	88% >=	10%
2544	Save Rate				26.09%	Stretch	22% >=	20%
2544	QA	81.68%				Target	80% >=	15%
2544	Adherence		74.34%			Below	85% <	0%
2544	Revenue			$335.40		Below	$750 <	0%

Thank you for you help.

[swuehl]

You probably have some code after the script snippet you've posted above, but just to be sure:

Are you dropping table load_IncentiveScores after transforming the columns to rows? If not, you will create a synthetic table that will link the columns with same name. In your scenario, I believe that's not what you want.

Do you have a synthetic table in your data model?

Again, if you want us to help you, you need to post more detailed information about your data model, best by creating a small sample QVW that demonstrates this issue.

[Anonymous]

App showing the Aggr() value on the wrong row.

I also attempted this with Rank() during my process but that is additional work on the load script.

Look at the table 'Using Aggr() - Minimum Score'  with EmpID 31480 if it's not selected.  You will see using Aggr( Min(AchievedIncentive), EmpID) places the lowest score in 15% when it's really 10%.  This works correctly on the majority of the values.

Thank you everyone for your help.  I have been able to create what I needed and part of the problem was how the app that feeds these numbers was built.  I have uploaded an app that takes the raw scores and builds what I need.  I will work further to have the back in the source app but this is working for now.