Basic company fuzzy matching between master and new sources

datawork

Hi all,

I am using https://www.youtube.com/watch?v=JlZiM2YSCCY as a guide to fuzzy match two source of company names. However, I am getting non-ideal outputs that I can use the community's help on to set me straight.

Attached is the self-contained flow in version 9 to illustrate. If you could make the change compatible in version 9 instead of 10, it would be great.

Obejctive is to match company names against each other. Master would have formal name like "Apple Inc."; new would have shortened name like Apple. I need the fuzzy match the two.

Master has:

Apple Inc.

Cisco Systems, Inc.

Microsoft Corporation

Oracle Corporation

New has:

Apple

Cisco

Microsoft

Oracle

After running the merger algorithm through with pre-defined "Company" matching. I only get 1 match instead of 4. Only Apple is being matched with Apple Inc.

Can you help me understand how to get to matching all 4 company between the 2?

Best,

Tom

Fuzzy matching training.yxzp

michael_treadwell

Tuning your Fuzzy Match tool to work with a data set is an iterative process. Many of the match functions you choose are not going to guarantee all of your fields match correctly. That being said, with the example data you gave, the fields in the New data set are wholly contained in the Master data set.

Next to Match Style in the Fuzzy Match configuration, choose Edit...

Under generate keys choose Whole Field. When Alteryx generates a key, it reads through the whole field and generates a key for each individual component of that field. Using 'Whole Field' will ensure that only keys where the components match will be output by Fuzzy Match.

As for Match Function, choose none.

This example output all four of the matches as you would expect but if your company names in New are not a subset of the company names in Master, this could present an issue.