>>

>>Jean d'Éléphant 
>>
>>JEANDELEPHANT

>>>JEANTELEPHANT
>>

In certain comparisons we go a step further, equalising D>>T, b>>p and some more,
if we have more than 1 column to compare/verify.

If you have ample time to run (read data cleaning at specific times) you might look into Levenshtein,
which is available in many dialects. Helps a lot when
"Mike's Mechanic Shop" is searched via "Mike Mechanic Shops" and similar cases.

regards

thomas