Spark2.3.1 Quick Start

Spark 2.3.1的编程指导内容如下：

Quick Start

Spark2.0前Spark主要编程接口是RDDs，2.0之后，RDDs接口被Dataset接口取代，Dataset和RDDs一样是强类型，但是底层做了大量优化。RDDs接口仍然可用，但是强烈推荐转换使用Dataset，性能比RDDs好

Spark Shell

基础

bin/spark-shell

Dataset操作

transformations

actions

缓存Caching

cache()是persist(StorageLevel.MEMORY_ONLY)的简写

persist(StorageLevel.MEMORY_ONLY)等

自包含应用

A self-contained application consists of a single, installable bundle that contains your application and a copy of the JRE needed to run the application. When the application is installed, it behaves the in the same way as any native application.