在dataframe上使用rapidfuzz

Question

我有4个列，它们是BuisnessID、Name、BuisnessID_y、Name_y和我希望用90%的相似性分数将名称与Name_y匹配，如果不是90%，则删除这些行。样本输入

df

BusinessID      NAME        BusinessID_y  NAME_y

1013120869  MANOJ WANKHADE  1013404164    SLIMI
1013120869  MANOJ WANKHADE  1013831688    AMOL SHAHAKAR
1013120869  MANOJ WANKHADE  1013376009    PRATHMESH AGRAWAL
1013120869  MANOJ WANKHADE  1013376009    PRATHMESH AGRAWAL
1013120869  MANOJ WANKHADE  1013478922    AMBRISH PANDRIKAR

我对python并不熟悉，也不知道如何做到这一点。另外，我有500 k的记录，所以任何其他快速毛绒的方法都会很棒。

Answer 1

>>> import pandas as pd
>>> import rapidfuzz
>>> df['matching_ratio'] = df.apply(lambda x:rapidfuzz.fuzz.ratio(x.NAME, x.NAME_y), axis=1).to_list()
>>> df

   BusinessID            NAME  BusinessID_y             NAME_y  matching_ratio
0  1013120869  MANOJ WANKHADE    1013404164              SLIMI       10.526316
1  1013120869  MANOJ WANKHADE    1013831688      AMOL SHAHAKAR       44.444444
2  1013120869  MANOJ WANKHADE    1013376009  PRATHMESH AGRAWAL       25.806452
3  1013120869  MANOJ WANKHADE    1013376009  PRATHMESH AGRAWAL       25.806452
4  1013120869  MANOJ WANKHADE    1013478922  AMBRISH PANDRIKAR       38.709677


>>> df[df.matching_ratio > 26] # change this '26' value to '90' as your requirmetn

   BusinessID            NAME  BusinessID_y             NAME_y  matching_ratio
1  1013120869  MANOJ WANKHADE    1013831688      AMOL SHAHAKAR       44.444444
4  1013120869  MANOJ WANKHADE    1013478922  AMBRISH PANDRIKAR       38.709677