Showing results for 
Search instead for 
Did you mean: 

Alteryx designer Discussions

Find answers, ask questions, and share expertise about Alteryx Designer.
Community v19.6

Looks aren't everything... But the latest Community refresh looks darn good!

Learn More

Association analysis for String fields - Identify hierarchies from a flat file


Hi All,



Imagine a dataset with several hierarchies potentially overlapping each others:

- Product / Product Group / Business Unit

- post code / city / country / region / continent

- customer / partner / distributor

- customer / sales rep

- customer / customer segment

- sales rep / manager



The challenge: 

The data is in a flat file hence we can't see these hierarchichal relationships (and we don't know what they should be)

Is there a way in Alteryx to analyze the data and identify correlation between string fields to rebuild these hierarchies?


Expected benefits:

- Recreate hierarchies easily in Excel / Tableau Output

- If output to database make it more efficient by creating several tables with relationships





Alteryx Partner

Sounds like an algorithm would be needed to do this efficiently, ideally already coded as an R package that you could run inside an Alteryx workflow. I did a few google searches, but couldn't find anything; sorry. Maybe with more time and diligence you could find something, versus trying to write something from scratch.


Looking at other tools I saw that MS Access attempt to do that with the table analyzer but I couldn't find much more.


Hey @Tibo - happy weekend!


Do you have some sample data that we can work with together?    If you're able to mock this up like @JoeM and @MattD do with the weekly challenges, then we can rope in the community to see if we can crack this together with you?


We do have challenges that sound similar, but if you have any example data then we can get a specific solution for your problem


Cheers @Tibo




Hi Sean,


Sorry for late reply!


I've enclosed the dataset. Really the challenge here would be to unflatten that data and identify the hierarchy automatically between columns.


The column names here should help you identify the ones that should be mapped but you'll note some inconsistancies.


By example  you'll see that for field [2. Range] values A4 and A6 are associated to two different values in [1. Techno] hence these two fields can't be linked in a direct hierarchy.