Fuzzy Match help

[ShreyanPuri]

Hi guys!

Facing an issue with fuzzy match (pasting sample data along with workflow below).
Despite the data being fairly similar, the match isn't giving me any results, even at a very low threshold (20%)

Was wondering what the reason behind this may be, and would be happy to hear any possible solutions.

(The ask is simply to map the 2 datasets against each other to get the closest possible match).

The input data has 2 datasets within it. I have flagged them as "market" or "company" data in field 2 (which acts as my source ID) and aim to compare results between these 2 and not within each of them, hence used the "merge" tool. The field I want to match is field 1: which contains the names.

What I've noticed is that alteryx seems to ignore matches below 85%, even if the threshold set is much lower. Don't know for sure

Thanks a ton in advance!

[mceleavey]

Hi @ShreyanPuri ,

You've used the wrong method in that you are only using one source. You therefore need to change the mode to "Purge Mode", which is poorly named.

Also, you can now change the Record ID field to literally be the field you're trying to match.

I've made the changes and now you get the following:

I hope this helps,

M

[ztriano]

Hi @ShreyanPuri I am not sure what you are trying to do here because you say you want to join two datasets, but you only have one dataset in here. But if you are trying to merge two datasets, I would reccomend the fuzzy join. It allows you to join two datasets with a match threshhold that you specify. It is a macro I found on the community page. You need to load the macro into your designer before you are able to do this yourself

To load the macro while it is open, right click in the whitespace, insert, macro, fuzzy join

But, if you like the macro, you can store it in a folder on your computer and link it to alteryx so you can easily use it whenever you want (and you can find it in your tool palete under Join). Macros | Alteryx Help

[ShreyanPuri]

Hi. Firstly, thanks!

I used the merged method because i wanted to compare the 2 different datasets (flagged under field 2 as either "market" or "company"). I want to compare names from between both these datasets, not within them hence i still believe the "merge" method is the correct way to go? 

The suggestion regarding record IDs was very helpful, though. Thanks!

[ShreyanPuri]

Thanks for posting the macro! Have edited the post to make the ask more clear. Will try and use to macro tool too!