I have been analyzing different options to perform a fuzzy search.
This feature is very useful when trying to consolidate entries with typos or simply different formatting preferences, so “Mr. Andrew Hill”, “Hill, Andrew R.” and “Andy Hill” all refer to the same underlying entity.
It seems that depending on the data source, the options for fuzzy search differ. For example:
-Named Entity Recognition module in Azure Machine Learning Studio is limited to people (PER), locations (LOC), and organizations (ORG) (see here for more information)
-Sharepoint Lists does not seem to have this feature available
So I was thinking a good feature for Microsoft Flow would be to offer a Fuzzy comparison action block in a similar way to the Excel add-in by taking data sources (or two arrays) as input. You can also select:
-the column pairs you want to compare (could be more than one)
-which columns from each source will be part of the output
-the similarity threshold
-the max number of matches
By having this feature as a Flow action block we could extend such feature to any data source.