As long as you’re filtering first and only then parsing the contents of notes, it shouldn’t cause any noticeable performance issues.
I regularly use Dataview tables based on the free-text contents of notes without any performance issues on desktop or mobile. This is from a vault with >20,000 notes.
As another example, my Air Quotes plugin can free-text filter the entire raw text of the bible as you type (750,000 words) with zero noticeable performance hit - you type and the results are immediate. I didn’t even bother to do any indexing - it’s just the regex function .matchAll(). Javascript/CPUs are pretty fast these days. A better developer could do it even more efficiently then I did.