I am trying to use Fuzzy match to be able to complete conflict checks for a company.
I've consolidated all 'Account Names' from both lists:
I only want to compare the 'Account Names' from different lists.
In the example below, I would like the output to match the POI source with account name 'Halcyon' to all Internal source account names that include 'Halcyon'
This is only 1 example where my account names are not hitting. I understand that the output will pick up multiple 'matches' that probably shouldn't be matched, but I would rather my final output be all inclusive and I can filter for actual matches afterwards.
any advice?
Solved! Go to Solution.
Hi @Mhynek1412 - I reviewed your workflow and applied a couple of changes:
1) Record ID Field in the Fuzzy Matching tool must be unique. In your WF you had duplicates. That's why I added a new column RecordIDFM.
2) I added the Join tool to remove perfect matches from the stream.
3) Key length 50 looks excessive. I reduced it to 4.
4) Given you're happy to pass more records and deal with them later I reduced the threshold to 65% to let these "Halcyon" records match. See below.
5) The Fuzzy Matching tool outputs duplicates and the Unique tool must be added after it. Alternative is to use the Make Group tool.
6) I added two Joins and one Union to gather the records together.
To conclude, Fuzzy Matching is not an exact science, it's an art. You may need to experiment to achieve the desired outcome. You can even apply it multiple times like below, to achieve the result. See example:
I attached adjusted workflow for your reference.
Thanks @ArtApa !
I've incorporated the changes into my WF. This helped to make my WF more sound but do you have any additional recommendations on how I can tweak the fuzzy match custom settings to get a bit more accuracy? I am seeing all the matches but the output is quite large.
Hi @Mhynek1412 - Watching this training may be helpful: https://community.alteryx.com/t5/Videos/Fuzzy-Matching-for-Beginners/td-p/330575