In my interpretation, food and drink are a subclass of objects. Creatures are not objects. If you kill your allies and they become corpses, then you could purify food and drink on the corpse. However, I would not allow the spell to work on the living creature.
Yes, this means that raw oysters might not be purifiable if still alive.
So, this is what I meant by an interesting ontological problem. Basically, whether or not this works, according to RAW, depends on how you categorize food. Clearly some food is eaten while alive. Probably, most food - killing it first is mostly reserved for animals. Some of the time. By some species. It is clear just looking at the real world that food is not really defined by any intrinsic quality within itself (i.e. being "an object," whatever arbitrary thing that means) but by its function in relationship to the eater: food is that which is eaten for nourishment (and drink that which is drunk). Don't believe me? Look up any definition of food that you want, and I promise you won't find the word "object" anywhere.
So it seems like the category of "food" is dependent on the biology and perceptions of the person casting the spell. Which could have all kinds of ramifications, especially in a fantasy game that allows all manner of fantastical creatures with fantastical diets to be characters. For example, what does a plasmoid categorize as food? For a Dhampir, food could be
dreams. Or what if the character casting the spell would not recognize the target as food - maybe a Thri-kreen has never encountered the strange food eaten by dwarfs and sees it as totally inedible - could they still cast the spell on it? Also, the "alive" argument is not so easily defined, either. For example, is an egg alive? What about a strawberry, straight off the vine?
I understand and agree with the argument that the spell would be overpowered if used this way, and it seems likely that it was never designed with this intent, especially since, as pointed out, higher level spells do the same thing, arguably less well. However, that is not my argument. My argument is that, RAW, and using normal, dictionary definitions of food and drink, I think that player has a valid point. And it's because of a kind of interesting epistemological problem inherent to the spell: it assumes that there is an agreed upon definition of "food and drink" but that is clearly not the case with D&D species (or even with real world species).