HomeLog InSource


[View] [Short] [Hash] [Raw]
So once we perform the intersection, what happens to the results?

Obviously we have to store them back to disk.

If we sort the tags, concatenate them, and then hash them, we could get a unique identifier to save the result under.

So our intersection method should accept a separate index to store the results in, and we should support incremental updates to it. All we have to do is check the last item and find it in each of our input arrays. Naturally, when we search for it, we should start at the end. If it's the last item in any of our inputs, then we don't need to do anything at all; it's already up to date.

It seems like this whole algorithm/data structure thing is actually shaping up pretty well, although it's taking longer than I'd like.