1)RDD五大特性
* - A list of partitions
每个RDD有一堆分区
* - A function for computing each split
对于RDD做计算,其实是对于每个分区做相同的function
* - A list of dependencies on other RDDs
RDD之类是有依赖关系的;RDDA -> RDDB -> RDDC -> RDDD 除了RDDA(加载),都有依赖关系
* - Optionally, a Partitioner for key-value RDDs (e.g. to say that the RDD is hash-partitioned)
(可选),默认hash的key-value RDD
* - Optionally, a list of preferred locations to compute each split on (e.g. block locations for an HDFS file)
(可选),计算每个分片的时候,会有最佳的位置 (locations 为何有s,因为HDFS来说,一个文件会有多个副本)