spark源码阅读一-spark读写文件代码分析

1.读取文件

从本地文件读取sparkcontext.textFile(“abc.txt”)

从hdfs文件读取sparkcontext.textFile("hdfs://s1:8020/user/hdfs/input”)

def textFile(

    path: String,

    minPartitions: Int = defaultMinPartitions): RDD[String] = withScope {

  assertNotStopped()

  hadoopFile(path, classOf[TextInputFormat], classOf[LongWritable], classOf[Text],

    minPartitions).map(pair => pair._2.toString).setName(path)

}

textFile第2个参数是partition个数，默认defaultMinPartitions为2，分2个区。也可以自定义最小分区数。

接下来就是看hadoopFile实现：

def hadoopFile[K, V](

    path: String,

    inputFormatClass: Class[_ <: InputFormat[K, V]],

    keyClass: Class[K],

    valueClass: Class[V],

    minPartitions: Int = defaultMinPartitions): RDD[(K, V)] = withScope {

  assertNotStopped()

  // This is a hack to enforce loading hdfs-site.xml.   

  // See SPARK-11227 for details.   

  FileSystem. getLocal(hadoopConfiguration)

  // A Hadoop configuration can be about 10 KB, which is pretty big, so broadcast it.   

  val confBroadcast = broadcast( new SerializableConfiguration(hadoopConfiguration))

  val setInputPathsFunc = (jobConf: JobConf) => FileInputFormat. setInputPaths(jobConf, path)

  new HadoopRDD(

    this,

    confBroadcast,

    Some(setInputPathsFunc),

    inputFormatClass,

    keyClass,

    valueClass,

    minPartitions).setName(path)

}

这里返回的就是HadoopRDD，RDD最关键的2个信息就是如何分区，以及PreferredLocation。

关于第一个问题分区，我们看看getPartitions

override def getPartitions: Array[Partition] = {

  val jobConf = getJobConf()

  // add the credentials here as this can be called before SparkContext initialized   SparkHadoopUtil. get.