我想计算PySpark DataFrame的两列之间的Jaro Winkler距离。 Jaro Winkler距离可通过所有节点上的pyjarowinkler软件包获得。
pyjarowinkler的工作方式如下:
from pyjarowinkler import distance
distance.get_jaro_distance("A", "A", winkler=True, scaling=0.1)
输出:
1.0
我正在尝试编写一个熊猫UDF以将两列作为Series传递,并使用lambda函数计算距离。这是我的操作方式:
@pandas_udf("float", PandasUDFType.SCALAR)
def get_distance(col1, col2):
import pandas as pd
distance_df = pd.DataFrame({'column_A': col1, 'column_B': col2})
distance_df['distance'] = distance_df.apply(lambda x: distance.get_jaro_distance(str(distance_df['column_A']), str(distance_df['column_B']), winkler = True, scaling = 0.1))
return distance_df['distance']
temp = temp.withColumn('jaro_distance', get_distance(temp.x, temp.x))
我应该能够在上述函数中传递任何两个字符串列。我得到以下输出:
+---+---+---+-------------+
| x| y| z|jaro_distance|
+---+---+---+-------------+
| A| 1| 2| null|
| B| 3| 4| null|
| C| 5| 6| null|
| D| 7| 8| null|
+---+---+---+-------------+
预期输出:
+---+---+---+-------------+
| x| y| z|jaro_distance|
+---+---+---+-------------+
| A| 1| 2| 1.0|
| B| 3| 4| 1.0|
| C| 5| 6| 1.0|
| D| 7| 8| 1.0|
+---+---+---+-------------+
我怀疑这可能是因为str(distance_df['column_A'])
不正确。它包含所有行值的串联字符串。
虽然此代码对我有用:
@pandas_udf("float", PandasUDFType.SCALAR)
def get_distance(col):
return col.apply(lambda x: distance.get_jaro_distance(x, "A", winkler = True, scaling = 0.1))
temp = temp.withColumn('jaro_distance', get_distance(temp.x))
输出:
+---+---+---+-------------+
| x| y| z|jaro_distance|
+---+---+---+-------------+
| A| 1| 2| 1.0|
| B| 3| 4| 0.0|
| C| 5| 6| 0.0|
| D| 7| 8| 0.0|
+---+---+---+-------------+
使用Pandas UDF可以做到这一点吗?我正在处理数百万条记录,因此UDF会很昂贵,但如果可以的话仍然可以接受。谢谢。
错误是由于您的函数在df.apply方法中引起的,请将其调整为以下值即可解决:
@pandas_udf("float", PandasUDFType.SCALAR)
def get_distance(col1, col2):
import pandas as pd
distance_df = pd.DataFrame({'column_A': col1, 'column_B': col2})
distance_df['distance'] = distance_df.apply(lambda x: distance.get_jaro_distance(x['column_A'], x['column_B'], winkler = True, scaling = 0.1), axis=1)
return distance_df['distance']
但是,Pandas df.apply方法没有向量化,这超出了我们在PySpark中在udf上需要pandas_udf的目的。一个更快,开销更少的解决方案是使用列表理解来创建返回的pd.Series(请检查link以获取有关Pandas df.apply及其替代品的更多讨论):
from pandas import Series
@pandas_udf("float", PandasUDFType.SCALAR)
def get_distance(col1, col2):
return Series([ distance.get_jaro_distance(c1, c2, winkler=True, scaling=0.1) for c1,c2 in zip(col1, col2) ])
df.withColumn('jaro_distance', get_distance('x', 'y')).show()
+---+---+---+-------------+
| x| y| z|jaro_distance|
+---+---+---+-------------+
| AB| 1B| 2| 0.67|
| BB| BB| 4| 1.0|
| CB| 5D| 6| 0.0|
| DB|B7F| 8| 0.61|
+---+---+---+-------------+