What for a question is that, "why not"? While this might sound funny, it's pretty normal in German:
Was ist das für eine Frage?
Was für eine Frage ist das?
This is also used with a sense of sarcasm for short
was für eine Frage?!
Was 'ne Frage.
I'd translate the former seeming halfway natural compared to other uses of for:
?What is that for a question
though obviously more frequently:
what a question
At any rate it's very frequent to hear
what did you do that for?
So the answer would be simple, what for already occupies a different semantics. Although wofür, wozu, warum, wieso [hast du das getan]? are almost equivalent in that specific example.
Regarding where-, the consequent answer is imaginably simple. Either (a) *hwar did not exclusively relate to location, or (b) reconstructing *hwar- for this compound is misguided. I'm leaning towards the latter, but this would require overturning the communis opinio, and seems impossible in light of whe-re, only motivated by Ger wofür, comparing Sp por que, Fr pour quoi; Lat quorsum ~ quo; ...
The accepted reconstruction is reasonable insofar "where" may be used for abstract concepts, where no explicit location is indicated. It alternates with when, just as Ger woher can replace the impersonal dative (why is there no wam) and a response might still use dem ("him, it, ..."? does not translate).
Ironically, glossing *hwar must be as polysemous as "where" itself, if the usage may be assumed to be conservative under (a). Ger wo can even be heared relating to people, "Der Mann, wo ..." (in my private experience) or time, "Der Moment, wo" (see ger.SE), it would be awesome if these conserved the unmarked PIE interrogative pronomina reconstructing *k^wo-, *k^we- (under (b)), but that is a difficult proposition to reconcile with old, exclusively written forms. In the same vein, Ger da would be worth a look.
"dar" brought in its conjunctive use, according to DWDS/W. Pfeifer as far back as attested. They give two etymologies for da, noting "dar", later merged with "da". Hence Da-tive--I'm kidding; the second da was tho and similar. dar- now only exists in compounds and correlatives, darstellen (display, produce), darum (therefore), etc., though chiefly dafür. It's maybe notable that rum, ran, raus, rein, etc. are considered ellipses of e.g. her-[um] (around) from the last millenium. That's a interesting shift for sure; especially so if positing that r- in that context actually reflects laryngeals, but nobody seriously believes this.
da is tentatively reconstructed for a Proto-Indo-European accusative *tam, and that at last is where it get's complicated. It looks as if it should correspond to ?wam or wem, dem (as noted above). For the locative, rather confere dort. The -r in dar is compared by Pfeifer to the masculine definite article der and that point one has to notice that location and a thing or person occupying that location are often equivalent, and that locations are often unknown and hence abstract. Thence it's understandable that "where" is yet acceptable for all these cases. herefore, I consider the question answered.
For an outlook I refer to [Dunkel, George E (2014), Lexikon der indogermanischen Partikeln und Pronominalstämme, Carl Winter Universitätsverlag], perhaps I should read it myself, but the outlook that many questions are still unanswered is daunting.. For an intensive study of Proto-Germanic syntax I sadly have nothing to offer. (I know this sounds cheap, and it puts this answer into perspective).