如何用字典值替换 pd.Series 列表值

问题描述

用NLTK词干后，得到字典：

> {'golden': 'gold','wonderfully': 'wonder','damaging': 'damag',> 'useless': 'use','toys': 'toy','ducks': 'duck'}

所以我想替换 pd.Series 中的现有单词（其中值是列表）：

0              ['want','buy','toys','ducks']
1                                    ['street']
2     ['damaging','isolating','toys']
3                       ['useless','clothing']

我希望输出为（replace玩具为玩具等......如字典中）：

0                ['want','toy','duck']
1                                    ['street']
2          ['damag','toy']
3                           ['use','clothing']

解决方法

尝试 explode 然后 replace 和 agg 回来，d 是你的字典

out = s.explode().replace(d).groupby(level=0).agg(list)

,

您可以使用 apply：

import pandas as pd

# the replacement dictionary
replacements = {'golden': 'gold','wonderfully': 'wonder','damaging': 'damag','useless': 'use','toys': 'toy','ducks': 'duck'}

# setup of the Series
s = pd.Series([['want','buy','toys','ducks'],['street'],['damaging','isolating','toys'],['useless','clothing']])


def replace(lst,repl):
    """This function receives a list and replaces the elements with the values of the keys in repl"""
    return [repl.get(e,e) for e in lst]


out = s.apply(replace,args=(replacements,))
print(out)

输出

0          [want,buy,toy,duck]
1                        [street]
2    [damag,isolating,toy]
3                 [use,clothing]
dtype: object