Because you would have to worry about deduplication of the results, which is nontrivial to do without incurring unbounded memory costs and/or messing up the expected ordering of the output. In short, it's not worth it.
Well, I'd expect no distinct ordering so the internal result could be piped to "sort --unique" to get rid of duplicates (I guess). If this (ordering) was documented, I assume that many users would still prefer running plocate instead of mlocate due to plocate's speed.
