(NFT standard) Why is the owner_id hashed as the prefix in tokens_per_owner?

[willemneal]

near-sdk-rs/near-contract-standards/src/non_fungible_token/core/core_impl.rs

Lines 362 to 371 in ddc5b14

	// Enumeration extension: Record tokens_per_owner for use with enumeration view methods.
	if let Some(tokens_per_owner) = &mut self.tokens_per_owner {
	let mut token_ids = tokens_per_owner.get(&owner_id).unwrap_or_else(|| {
	UnorderedSet::new(StorageKey::TokensPerOwner {
	account_hash: env::sha256(owner_id.as_bytes()),
	})
	});
	token_ids.insert(&token_id);
	tokens_per_owner.insert(&owner_id, &token_ids);
	}

It seems that you end up paying more to 1) hash the string and 2) store more bytes than with just the account_id.

[austinabell]

It seems that you end up paying more to 1) hash the string and 2) store more bytes than with just the account_id.

Depends on the size of the account id. Account Ids can be up to 64 bytes in length, whereas the hash is always 32. I'm not sure if this change was benchmarked, but possibly was done to close the gap in costs between long and short lengths?

This doesn't affect the interface so it's really not part of the standard and if for whatever reason someone wanted to modify this, they would have the freedom to.