Python pyspark.ml管道:基本预处理任务是否需要自定义转换器?
从Python pyspark.ml管道:基本预处理任务是否需要自定义转换器?,python,apache-spark,machine-learning,pyspark,data-science,Python,Apache Spark,Machine Learning,Pyspark,Data Science,从pyspark.ml和管道API开始,我发现自己正在为典型的预处理任务编写自定义转换器,以便在管道中使用它们。示例: from pyspark.ml import Pipeline, Transformer class CustomTransformer(Transformer): # lazy workaround - a transformer needs to have these attributes _defaultParamMap = dict() _p
pyspark.ml
和管道API开始,我发现自己正在为典型的预处理任务编写自定义转换器,以便在管道中使用它们。示例:
from pyspark.ml import Pipeline, Transformer
class CustomTransformer(Transformer):
# lazy workaround - a transformer needs to have these attributes
_defaultParamMap = dict()
_paramMap = dict()
_params = dict()
class ColumnSelector(CustomTransformer):
"""Transformer that selects a subset of columns
- to be used as pipeline stage"""
def __init__(self, columns):
self.columns = columns
def _transform(self, data):
return data.select(self.columns)
class ColumnRenamer(CustomTransformer):
"""Transformer renames one column"""
def __init__(self, rename):
self.rename = rename
def _transform(self, data):
(colNameBefore, colNameAfter) = self.rename
return data.withColumnRenamed(colNameBefore, colNameAfter)
class NaDropper(CustomTransformer):
"""
Drops rows with at least one not-a-number element
"""
def __init__(self, cols=None):
self.cols = cols
def _transform(self, data):
dataAfterDrop = data.dropna(subset=self.cols)
return dataAfterDrop
class ColumnCaster(CustomTransformer):
def __init__(self, col, toType):
self.col = col
self.toType = toType
def _transform(self, data):
return data.withColumn(self.col, data[self.col].cast(self.toType))
它们可以工作,但我想知道这是一种模式还是反模式——这样的转换器是使用管道API的好方法吗?是否有必要实现它们,或者其他地方是否提供了等效的功能?我想说,它主要是基于观点的,尽管它看起来不必要地冗长,而且Python
Transformers
与管道
API的其余部分没有很好地集成
还值得指出的是,这里的所有内容都可以通过SQLTransformer
轻松实现。例如:
from pyspark.ml.feature import SQLTransformer
def column_selector(columns):
return SQLTransformer(
statement="SELECT {} FROM __THIS__".format(", ".join(columns))
)
或
只要稍加努力,您就可以使用SQLAlchemy和Hive方言来避免手写SQL。您能否详细说明一下“Python转换器与管道API的其余部分没有很好地集成”?默认情况下,没有
mlwriteable
(尽管有)。SQL并不是我所希望的优雅的替代方案,但是答案很好->接受。你怎么称呼定制变压器?
def na_dropper(columns):
return SQLTransformer(
statement="SELECT * FROM __THIS__ WHERE {}".format(
" AND ".join(["{} IS NOT NULL".format(x) for x in columns])
)
)