如何移除Pandas数据框列中每个单词末尾的‘br’字母组合？

问题背景 #

需要处理Pandas数据框中的某一列（每行都是独立句子），移除每个单词末尾的"br"字母——这些是之前清理HTML内容时未妥善处理<br>标签导致的无效后缀，比如出现somebr、andbr这类错误词汇。

错误示例 #

Sentence = here are somebr examples of poorly written paragraphs andbr well-written paragraphsbr on the same topicbr how do they compare?

期望结果 #

Sentence: here are some examples of poorly written and well-written paragraphs on the same topic how do they compare?

核心要求 #

• 仅移除单词末尾的"br"，不影响句子结构和其他正常单词
• 保留brutish、breathtaking、ember这类中间含"br"的正常词汇
• 无需要保留的末尾为"br"的单词

解决方案 #

使用Pandas的str.replace()方法配合正则表达式，精准匹配单词末尾的"br"并替换为空，避免误改正常词汇。

代码实现 #

import pandas as pd

# 假设数据框为df，目标列名为'Sentence'
df['Sentence'] = df['Sentence'].str.replace(r'(\w+)br\b', r'\1', regex=True)

正则逻辑说明 #

• (\w+)：捕获单词的主体部分（一个或多个字母/数字/下划线），通过括号分组便于后续引用
• br：精准匹配单词末尾的"br"后缀
• \b：单词边界标记，确保"br"处于单词结尾位置，不会匹配到brutish这类中间含"br"的单词
• \1：替换为之前捕获的单词主体，即去掉末尾"br"后的内容

测试验证 #

用示例句子测试：

test_str = "here are somebr examples of poorly written paragraphs andbr well-written paragraphsbr on the same topicbr how do they compare? Also, brutish and breathtaking should stay unchanged."
cleaned_str = pd.Series([test_str]).str.replace(r'(\w+)br\b', r'\1', regex=True)[0]
print(cleaned_str)

输出结果：

here are some examples of poorly written paragraphs and well-written paragraphs on the same topic how do they compare? Also, brutish and breathtaking should stay unchanged.

可见错误后缀被移除，正常含"br"的单词未受影响，完全符合需求。

内容的提问来源于stack exchange，提问作者user15516822