I have been taking a class where we primarily use JMP. JMP has a nice feature where you can ‘recode’ a column. For example let’s say you have a bunch of patients of a clinic the same person has had their name typed in differently several times as Michael Becker, another time as Michael s. becker, and another couple times as Michael s Becker. It’s really easy to fix it by replacing them all with ‘Michael S. Becker’ so they all match. JMP will also show all the different instances so it’s easy to find and fix as well. Does Knime have that capability? I hope I’m explaining it well enough to make sense.
do you know for each name which patient it really belongs to? If you have a patient ID or something like that, you can use GroupBy with group column ID and aggregation method “mode” on the name to find the most used name per patient, then use the Joiner using the patient ID to join back to the original table. We also have things like Similarity Search to help you when an ID is not available and you simply want to compare strings with each other. However, there is no one-node-solution to this, I’m afraid.
This topic was automatically closed 182 days after the last reply. New replies are no longer allowed.