Figuring out whether two entities (people, places, products, etc.) are the same can always be a challenge, particularly when dealing with messy data with no unique identifier. You can’t ignore the problem though, because it can be crucial to joining datasets from different sources or de-duplicating a dataset. Similarly, categorizing and transforming messy data can also be a challenge and require complex logic that is hard to manage or even impossible to define.

EntityMatcher is an application that let's you do these tasks easily and automatically with AI. Simply provide the datasets and EntityMatcher will do the rest. It also provides a UI for reviewing the matches yourself. This add-on allows users to do this all without having to leave sheets by using custom formulas.

For matching, use the ENTITYMATCH formula, where you specify the datasets as ranges, the id columns, and the job name:
=ENTITYMATCH(A1:E51,"id", google!A1:E44, "id", "test_google_sheets_matching")

For transformation, use the ENTITYTRANSFORM formula, where you specify the dataset, id column, instructions for the transformation that needs to take place, and the job name:
=ENTITYTRANSFORM(A1:E44, "id", "get the vendor and indicate whether the product is software or not, two separate fields", "plugin-transform-test-v1")

For categorization, use the ENTITYCATEGORIZE formula, where you specify the dataset, id column, categories, and the job name:
=ENTITYCATEGORIZE(A1:E44, "id", I2:I3, "plugin-categorize-test-v1")

All of these formulas use the EntityMatcher API to run AI models to automatically complete the relevant task. The outputs are displayed directly in the google sheet, and you can use the EntityMatcher UI to review the predictions in more detail if necessary.

Check out https://docs.entitymatcher.com/quickstart for more information.