Fuzzy matching not as significantly effective

[AA007]

i have apple and banana;banana and apple in a same column with different count in a different column. It is only a sample of the data. The data is dynamic. Is there a way I could check for strings that are symmetric or contain same number of strings. I even tried with text to columns but the string is not limited to two or three number of words. will tokenizing it and checking reverse work?

[PhilippK]

Hi @AA007 ,

I think fuzzy matching could help you here.

Can you share sample data with us?

Fuzzy Matching requires some configuration iterations and thus without data it is hard to provide you with a suitable solution.

Best regards

Phil

[AA007]

new_SourceTitle Sum_Count Right_new_SourceTitle Right_Sum_Count
CFO & Secretary 1 CFO Secretary 160
CFO Secretary & Treasurer 53 CFO Secretary Treasurer 13
CFO & Senior VP of Administration 1 CFO Senior VP Administration 2
CFO & Treasurer 1 CFO Treasurer 456
CFO Treasurer & Secretary 68 CFO Treasurer Secretary 4
CFO Treasurer & VP 3 CFO Treasurer VP 2
CFO & VP of Administration 4 CFO VP Administration 6
CFO & VP of Finance 40 CFO VP Finance 17
CFO & VP Finance & Operations 1 CFO VP Finance & Operations 1

[AA007]

Agent & Broker	3	Agent Broker	Broker Agent	5
Owner Founder & President	19	Owner Founder President	Owner President & Founder	17
Founder President & CEO	69	Founder President CEO	Founder President CEO	66
VP Finance	69	Finance VP	Finance & VP	67
Founder CEO & Chairman	22	Founder CEO Chairman	Founder CEO Chairman	23
Founder Chairman & CEO	33	Founder Chairman CEO	Chairman CEO Founder	33

[AA007]

at the end i want to be able to keep the first column but the next column of data has the same data and be able to keep the count and add up through a summarize table

[PhilippK · Accepted solution]

Hi @AA007 ,

please find the solution attached.

Not so sure whether the final output has the exact column structure you are looking for, but you can easily adjust this in the workflow:   

Best regards

Phil

[ArtApa · Accepted solution]

Hi @AA007 - Please be aware that by definition Fuzzy Matching attempts to find a match which, although not a 100 percent match, and is above the threshold matching percentage set by the application. Fuzzy Matching is an art and is really good in the following situations:

•Names / nicknames

•Addresses

•Company names

•Human error / typos

Alternatively, you can always try to find exact matches before applying Fuzzy Matching. If your data is always like the provided sample then you can use something like this: