There's loads of implementations of a BigMap on npm, but I didn't really like any of them.

Well, honestly, I just wanted to make my own for the fun of it.

https://www.npmjs.com/package/bigbigmapset

https://github.com/JesseRussell411/BigBigMapSet

There's a BigMap and BigSet.

Here's why I like my implementation:

• typescript support

• runtime independent (theoretically) so long as Map and Set throw a RangeError when they're full.

  • From experimentation, it looks like chrome's limit is actually 2^23 instead of node's 2^24, but my implementation just waits until a RangeError is thrown to allocate a new chunk so it doesn't matter what the size limit is (other errors passed to the caller).

• They actually extend Map and Set so instanceof works and typescript won't complain if you use them where a Map or Set were expected. Thanks to extending Map and Set, they ARE Map and Set.

• They exactly copy Map and Set's interface, including giving the constructor an Iterable of entries. You just add "Big" to the constructor call in your source code and the rest should just work.

  • They maintain insertion order.

• probably relatively performant. There aren't any lambdas, forEach()s, map()s, filter()s, etc in the source code because they tend to be slow and this is the kind of thing that needs to be as fast as possible. Of course, it's still a very simple implementation that just stores an un-indexed list of maps or sets and searches through them linearly for each operation.

  • There's an implementation as part of Keyv that actually indexes the chunks which looks interesting; though I can't imagine it makes much of a positive difference until you have more than a few chunks (like over 32): https://github.com/jaredwray/keyv/blob/main/packages/bigmap/src/index.ts