allTokensIndex

allTokensIndex is a storage mapping used in enumerable token contracts to track the position of each token in the allTokens array. In typical ERC721Enumerable-like implementations, the contract maintains an array called _allTokens containing every tokenId currently in existence, and a mapping called _allTokensIndex that maps each tokenId to its index within that array. This pairing enables constant-time lookups of a token’s position and supports enumeration of all tokens via functions like tokenByIndex and totalSupply.

How it works: When a new token is minted, its tokenId is appended to the _allTokens array and _allTokensIndex[tokenId] is set to the new last index (length minus one). When a token is burned, the usual pattern is to swap the token at the burned token’s index with the last token in the array, update the swapped token’s index in _allTokensIndex, remove the last element from _allTokens, and delete the entry for the burned token from _allTokensIndex. Transfers between owners do not affect the _allTokens array, so the index of a given tokenId remains stable across transfers.

Typical usage: tokenByIndex(uint256 index) returns the tokenId at _allTokens[index], and totalSupply() returns the length of _allTokens. The _allTokensIndex mapping is usually private and maintained by internal functions that manage the enumeration state.

Considerations: Implementing allTokensIndex enables full token enumeration at the cost of additional storage and gas. It is a convention associated with ERC721Enumerable and similar patterns; exact naming may vary across implementations.