Massive data topn problem

1. Find the top ten thousand 100？

answer ： Quick line up patition+ Two points , Pile up

2. 10 The figure of 100 million is the top 10 ten thousand , Space for 1 Billion ？

answer ：（ Divide and conquer plus stack ）, Time complexity klogk

3. 100 Billion figures to find the top 10 Billion , Space 1 Billion ？

answer ： This problem is not an algorithm problem , It should be the realization of the project , Basically, it depends on slicing to multiple machines for parallel operation , How to do it , Look at that 64 The horse looks for the top four problems .

All in all , For massive data , It's all for “map-reduce” To divide and rule , Separate local treatment , Then sum it up .

Reference resources ：

In massive data TOPK Summary of problems - Excaliburer - Blog Garden

Massive data processing top K problem - Be happy - Blog Garden

In massive data TopK problem _ A blog with eyes reflecting the stars -CSDN Blog _ Huge amounts of data topk problem