如何在 spark RDD (JavaRDD) 中获取记录的文件名

Question

我正在使用将多个文件加载到 JavaRDD 中

JavaRDD<String> allLines = sc.textFile(hdfs://path/*.csv);

加载文件后，我修改了每条记录并想保存它们。但是，我还需要将原始文件名 (ID) 与记录一起保存以供将来参考。无论如何，我可以从 RDD 中的单个记录中获取原始文件名吗？ 谢谢

Answer 1

您可以尝试在以下 sn-p 中执行类似操作：

JavaPairRDD<LongWritable, Text> javaPairRDD = sc.newAPIHadoopFile(
    "hdfs://path/*.csv", 
    TextInputFormat.class, 
    LongWritable.class, 
    Text.class, 
    new Configuration()
);
JavaNewHadoopRDD<LongWritable, Text> hadoopRDD = (JavaNewHadoopRDD) javaPairRDD;

JavaRDD<Tuple2<String, String>> namedLinesRDD = hadoopRDD.mapPartitionsWithInputSplit((inputSplit, lines) -> {
    FileSplit fileSplit = (FileSplit) inputSplit;
    String fileName = fileSplit.getPath().getName();

    Stream<Tuple2<String, String>> stream =
        StreamSupport.stream(Spliterators.spliteratorUnknownSize(lines, Spliterator.ORDERED), false)
            .map(line -> {
                String lineText = line._2().toString();
                // emit file name as key and line as a value
                return new Tuple2(fileName, lineText);
            });
    return stream.iterator();
}, true);

更新（针对 java7）

JavaRDD<Tuple2<String, String>> namedLinesRDD = hadoopRDD.mapPartitionsWithInputSplit(
    new Function2<InputSplit, Iterator<Tuple2<LongWritable, Text>>, Iterator<Tuple2<String, String>>>() {
        @Override
        public Iterator<Tuple2<String, String>> call(InputSplit inputSplit, final Iterator<Tuple2<LongWritable, Text>> lines) throws Exception {
            FileSplit fileSplit = (FileSplit) inputSplit;
            final String fileName = fileSplit.getPath().getName();
            return new Iterator<Tuple2<String, String>>() {
                @Override
                public boolean hasNext() {
                    return lines.hasNext();
                }
                @Override
                public Tuple2<String, String> next() {
                    Tuple2<LongWritable, Text> entry = lines.next();
                    return new Tuple2<String, String>(fileName, entry._2().toString());
                }
            };
        }
    }, 
    true
);

Answer 2

你想要 spark 的 wholeTextFiles 函数。来自文档：

For example, if you have the following files:

   hdfs://a-hdfs-path/part-00000
   hdfs://a-hdfs-path/part-00001
   ...
   hdfs://a-hdfs-path/part-nnnnn

Do val rdd = sparkContext.wholeTextFile("hdfs://a-hdfs-path"),

then rdd contains

   (a-hdfs-path/part-00000, its content)
   (a-hdfs-path/part-00001, its content)
   ...
   (a-hdfs-path/part-nnnnn, its content)

它返回一个元组的 RDD，其中左边是文件名，右边是内容。

Answer 3

您应该可以使用toDebugString。使用 wholeTextFile 会将文件的全部内容作为一个元素读入，而 sc.textfile 会创建一个 RDD，每行作为一个单独的元素 - 如 here 所述。

例如：

val file= sc.textFile("/user/user01/whatever.txt").cache()

val wordcount = file.flatMap(line => line.split(" ")).map(word => (word, 1)).reduceByKey(_ + _)

wordcount.toDebugString

// res0: String =

// (2) ShuffledRDD[4] at reduceByKey at <console>:23 []

// +-(2) MapPartitionsRDD[3] at map at <console>:23 []

//    |  MapPartitionsRDD[2] at flatMap at <console>:23 []

//    |  /user/user01/whatever.txt MapPartitionsRDD[1] at textFile at <console>:21 []

//    |  /user/user01/whatever.txt HadoopRDD[0] at textFile at <console>:21 []