I often use FME to classify data.

One of the ways I do this: Aggregate the data on a text field, use this as input for a lookup table, create the lookup table in Excel, read the Excel with FME and join the lookup table with the original data using a FeatureMerger. This works perfectly but it's static, so when the data changes and new variations arise I need to update the lookup table.

The alternative is to use an AttributeCreator with a Conditional value, operator Contains and build a more dynamic lookup. But this gets complex very quickly because of the order of rules. And it gives only the first result. This way it is hard to check for mistakes. And I really like to work with the lookup table in Excel.

It would be great if the FeatureMerger could do this: Merge when de requester attribute contains the supplier attribute. (add operator next to Comparison Mode) Process duplicate suppliers create a list. I understand performance challenges when comparing large data vs large lookup tables.

I can create a custom transformer with a StringSearcher looping through the lookup table but I wondered if someone encountered the same and found a smarter route?