Question:
Massive sorting data: Where to
put database, no collection class (memory overflow), is stored in a file,
limited memory, 50 or 50 billion words digital text to sort?
Stable Sort & Unstable Sort:
https://www.cnblogs.com/codingmylife/archive/2012/10/21/2732980.html
Solve the problem
Multi- threaded high concurrency, massive data sorting, query, statistics
C10K-> refers to The server supports the problem of thousands of clients at the same time, that is, concurrent 10 000 connections (this is also the origin of the name c10k).
Solution:
Multi-pass merge sort:
1 file is split into 100 files, merge sort is performed between multiple files, and finally large files are merged.
Idea: divide and conquer
java.io and java.
internal sorting, external sorting
Is there any other way:
haddoop underlying bitmap sorting method
mapReduce