Find and Replace unsuccessful due to messy Data

Question

Hey everybody,

I need to match two inputs that have an identical build and should have the same content in the rows of the columns I'm trying to match at.

What I realized, though, is that the creator of the tables must have been careless since there are issues such as missing letters (e.g. New Yor instead of New York) in one of the two tables.

Is there any way to clean that up in the WF?

Thank you for your answers!

RobbieA · Answer

I'm unfamiliar with fuzzy match, if you have a small amount of rows to correct it might be worth changing them manually. Let me know if it works. If you post a sample of the problem, I'm happy to have a go at solving it.

Cela · Answer

@RobbieA  The affected area of the workflow is really just the two input tools as well as the find and replace tool, no other tools are being used before, the ones after shouldn't be affecting that part.

So it really is a simple build at this point.

I am experimenting with fuzzy match but have no clue how to properly use it yet as I haven't had the opportunity or necessity to work with it before. Do you think that might be a way to match them based on the Match Threshold?

RobbieA · Answer

You can join the data sets and observe which rows are misspelt from the L/R anchor (therefore don't join). From there you can clean up your rows with the Replace() function in the formula tool or regex, before adding it back to your workflow (Or add this step in before the join).

Do you have an example of the workflow?