pandas：只替换单词而不是整个句子

Posted 2023-02-23

【中文标题】pandas：只替换单词而不是整个句子

【英文标题】：pandas: replace only the word and not the entire sentence

【发布时间】：2022-01-13 09:04:19

【问题描述】：

我有一个如下的数据框：(e,g)

import pandas as pd
df = pd.DataFrame('text':['Lary Page is visiting on Saturday',' On Monday his boss, Maria Jackson is here .'])

我想将以下列表中收集的星期几替换为来自假库中的随机日期，用于每次出现的日子。我做了以下事情：

from faker import Faker
import numpy as np
fake = Faker()

days_list = ['Saturday','Monday','Tuesday']

我尝试了以下方法，但都返回了替换的日期而不是整个句子：

df.text = np.where(df.text.str.contains('|'.join(days_list)),
               fake.day_of_week(), df.text)

或

df.text.str.replace('|'.join(days_list), fake.day_of_week())

我想要的输出：

print(df): (e,g)
'Lary Page is visiting on Tuesday'
'On Thursday his boss, Maria Jackson is here .'

【参考方案1】：

使用 lambda 函数替换回调：

regex = '|'.join(days_list)
df['text'] = df.text.str.replace(regex, lambda x: fake.day_of_week(), regex=True)
print (df)
                                             text
0                Lary Page is visiting on Tuesday
1   On Thursday his boss, Maria Jackson is here .

【参考方案2】：

from faker import Faker
import pandas as pd
df = pd.DataFrame('text':['Lary Page is visiting on Saturday',' On Monday his boss, Maria Jackson is here .'])

fake = Faker()
days_list = ['Saturday','Monday','Tuesday']
df['text'] = df['text'].apply(lambda x: ' '.join(fake.day_of_week() if i in days_list else i for i in x.split()))

print(df)

输出：

                                          text
0             Lary Page is visiting on Tuesday
1  On Monday his boss, Maria Jackson is here .