Recently I found that I’m kind of over talking when I speak with others, so I want to train my conclusion of solving problem. Here are some problems I meet when I read books of Big Data.
- No 1. Why
Sparkdoes not use map-reduce?
Disk reading is too slow to complete big data analyze.Map-reduce is a shuffle which meanings parallizing.Spark just use memory to complete parallizing a huge problem.
- No 2. How Hadoop to append data?
Just use