如何从Spark Dataframe中的Row对象中获取值

Question

for

averageCount = (wordCountsDF
                .groupBy().mean()).head()

我

行(avg(数)= 1.6666666666666667)

但当我尝试:

averageCount = (wordCountsDF
                .groupBy().mean()).head().getFloat(0)

我得到以下错误:

AttributeError: getFloat--------------------------------------------------------------------------- AttributeError 回溯(最近调用in () in ()1 #待办事项:用合适的代码替换----> 2 averageCount = (wordCountsDF .;3 . .mean groupBy () ()) .head () .getFloat (0)45 print averageCount

/databricks/spark/python/pyspark/sql/types.py in getattr

除了ValueError:-> 1272 raise AttributeError(item) 1273 1274 def setattr(self, key, value):

AttributeError: getFloat

我做错了什么?

Answer 1

我明白了。这将返回值:

averageCount = (wordCountsDF
                .groupBy().mean()).head()[0]

Answer 2

这也可以:

averageCount = (wordCountsDF
                .groupBy().mean('count').collect())[0][0]
print averageCount

Answer 3

数据框行继承自namedtuple(来自collections库)，因此尽管您可以像上面那样对传统元组进行索引，但您可能希望通过其字段的名称访问它。毕竟，这就是命名元组的意义所在，而且它对于将来的更改也更加健壮。这样的:

averageCount = wordCountsDF.groupBy().mean().head()['avg(jobs)']