检查类型：如何检查某些东西是RDD还是DataFrame？

检查类型：如何检查某些东西是RDD还是DataFrame？

isinstance

会很好地工作：

from pyspark.sql import Dataframefrom pyspark.rdd import RDDdef foo(x):    if isinstance(x, RDD):        return "RDD"    if isinstance(x, Dataframe):        return "Dataframe"foo(sc.parallelize([]))## 'RDD'foo(sc.parallelize([("foo", 1)]).toDF())## 'Dataframe'

但是单次调度是一种更为优雅的方法：

from functools import singledispatch@singledispatchdef bar(x):    pass@bar.register(RDD)def _(arg):    return "RDD"@bar.register(Dataframe)def _(arg):    return "Dataframe"bar(sc.parallelize([]))## 'RDD'bar(sc.parallelize([("foo", 1)]).toDF())## 'Dataframe'

如果您不介意其他依赖项

multipledispatch

也是一个有趣的选择：

from multipledispatch import dispatch@dispatch(RDD)def baz(x):    return "RDD"@dispatch(Dataframe)def baz(x):    return "Dataframe"baz(sc.parallelize([]))## 'RDD'baz(sc.parallelize([("foo", 1)]).toDF())## 'Dataframe'

最后，最Python化的方法是简单地检查一个接口：

def foobar(x):    if hasattr(x, "rdd"):        ## It is a Dataframe    else:        ## It (probably) is a RDD