[Tutor] Any ideas on designing an efficient algorithm for this problem?

[Mats Wichmann]

On 6/24/23 19:50, ThreeBlindQuarks via Tutor wrote:
> 
> I am wondering if a sliding window protocol is a suitable way to find the top 100.

That's not a bad approach.

There's also an algorithm already in the standard library: 
heapq.nlargest(100, array)

whether that's optimal for a billion records, I don't know, but the heap 
sorting approach is usually considered pretty efficient.