IndexError：在pyspark shell上使用reduceByKey操作时列出索引超出范围

Question

目标：从 youtube 数据集中找到最高的视频类别

使用：Pyspark 外壳

预期：类别及其出现次数

实际：使用 reduceBykey 作为 IndexError 时出错：列表索引超出范围

我试过下面的代码：

data="/Users/sk/Documents/GitRepository/Udemy_BigData_spark/1.txt"
input = sc.textFile(data)
results = input.map(lambda x: (x.split(‘\t')[3].encode("utf-8").replace('"', '').replace("'", '')))results.take(20)

这给出了以下结果：

['Comedy', 'Comedy', 'Entertainment', 'People & Blogs', 'People &
Blogs', 'Music', 'Comedy', 'People & Blogs', 'Entertainment',
'Entertainment', 'Entertainment', 'Entertainment', 'Entertainment',
'Entertainment', 'Entertainment', 'Entertainment', 'Entertainment',
'Entertainment', 'Entertainment', 'Entertainment']


results=results.map(lambda x: (x,1))

这给出了以下结果：

[('Comedy', 1), ('Comedy', 1), ('Entertainment', 1), ('People & Blogs', 1), ('People & Blogs', 1), ('Music', 1), ('Comedy', 1), ('People & Blogs', 1), ('Entertainment', 1), ('Entertainment', 1), ('Entertainment', 1), ('Entertainment', 1), ('Entertainment', 1), ('Entertainment', 1), ('Entertainment', 1), ('Entertainment', 1), ('Entertainment', 1), ('Entertainment', 1), ('Entertainment', 1), ('Entertainment', 1)]


results=results.reduceByKey(lambda x, y: x + y)    
results.take(20)

这给出了一个巨大的错误:(

我希望它显示如下结果：

(179049,Music), (127674,Entertainment), (87818,Comedy), (73293,Film &
Animation), (67329,Sports)

Answer 1

我写的代码是在scala中的；

val ds = Seq("A", "B", "C", "A", "B", "C", "D", 
             "E", "F", "G", "A")
  .toDF.as[String].map(x => (x, 1))
ds.groupByKey(x => x._1)
  .reduceGroups((l, r) => (l._1, l._2+r._2))
  .show

输出：

+-----+------------------------------+
|value|ReduceAggregator(scala.Tuple2)|
+-----+------------------------------+
|    F|                        [F, 1]|
|    E|                        [E, 1]|
|    B|                        [B, 2]|
|    D|                        [D, 1]|
|    C|                        [C, 2]|
|    A|                        [A, 3]|
|    G|                        [G, 1]|
+-----+------------------------------+