I figured out a solution to my problem. Not using the logic I had mentioned in my original post, but by using a workaround (or maybe this is the best way?). This is what I did:
Instead of removing the fruits that I don't care about, I split them out into their own rows using SubField and then cleaned up the column to remove extra spaces, using LTrim, so the data comes in clean row-by-row.
Added this into my loadscript:
LTrim(SubField([Types of Fruit], ',')) as [Fruit]
Ideally, I'd love to be able to remove everything I don't want at the loadscript level. Still working on that!
Using SubField() is a great way if you want them split eventually anyways. But if did not want them split, WildMatch should work like this:
LOAD * INLINE [
Apple, Pear, Peach
Banana, Orange, Apple
Peach, Tomato, Kiwi
Apple, Orange, Peach
] (delimiter is '|')
WHERE WildMatch(Fruit, '*Apple*', '*Banana*')