从文本提取多格式日期：DataFrame日期提取正则问题求助

你当前代码的核心问题： #

1. str.extract使用错误：str.extract仅返回第一个匹配的分组结果，而非所有匹配项，且会将正则中的分组拆分成多列，无法得到你需要的所有日期拼接字符串。
2. 正则模式覆盖不全：
   • 未匹配月份全称（如June、September）和大小写变体（如DEC、APRIL）
   • 未处理用to连接的日期格式（如03/2014 to 08/2015）
   • 未兼容仅含年份的格式（如(2015 – 2016)）
   • 未考虑括号包裹的场景
3. 正则结构冗余：过多的可选分组导致匹配逻辑混乱，无法精准覆盖目标日期格式。

修正后的实现代码 #

import pandas as pd

def extract_dates(df, column):
    # 正则模式：覆盖所有目标日期格式，忽略大小写
    pattern = r'(?i)\(?\s*(?:(?:Jan(?:uary)?|Feb(?:ruary)?|Mar(?:ch)?|Apr(?:il)?|May|Jun(?:e)?|Jul(?:y)?|Aug(?:ust)?|Sep(?:tember)?|Oct(?:ober)?|Nov(?:ember)?|Dec(?:ember)?)[\s,-]*\d{2,4}|\d{1,2}/\d{2,4}|\d{2,4})\s*(?:[-–]|to)\s*(?:(?:Jan(?:uary)?|Feb(?:ruary)?|Mar(?:ch)?|Apr(?:il)?|May|Jun(?:e)?|Jul(?:y)?|Aug(?:ust)?|Sep(?:tember)?|Oct(?:ober)?|Nov(?:ember)?|Dec(?:ember)?)[\s,-]*\d{2,4}|\d{1,2}/\d{2,4}|\d{2,4})\s*\)?'
    
    # 提取所有匹配的日期，并用逗号拼接
    df['Dates'] = df[column].str.findall(pattern).str.join(', ')
    return df

# 测试示例
sample_text = """Sales Assistant @ DFS Duration - June 2021 - 2023 Currently working in XYZ Within the role I am expected to achieve sales targets which I currently have no problems reaching. Job Role/Establishment - Plasterer @ XX Plasterer’s Duration - September 2016 - Nov 2016 Job Role/Establishment - Customer Advisor @ AA Duration - (2015 – 2016) Job Role/Establishment - Warehouse Operative @ xyz Duration - 03/2014 to 08/2015 In the xyz warehouse  Job Role/Establishment - Airport Terminal Assistant @ port Duration - 01/2012 - 06/2013 Working at the airport . Job Role/Establishment - Apprentice Floorer @ YY Floors Duration - DEC 2010 – APRIL 2012"""

df = pd.DataFrame({'id': ['01'], 'Raw_text': [sample_text]})
df = extract_dates(df, 'Raw_text')
print(df[['id', 'Raw_text', 'Dates']])

正则模式说明 #

• (?i)：忽略大小写，匹配DEC/dec/December等变体
• \(?/\)?：匹配可选的前后括号
• 核心匹配单元：
  • 月份+年份：(?:Jan(?:uary)?|...|Dec(?:ember)?)[\s,-]*\d{2,4}（支持缩写/全称）
  • 数字月份+年份：\d{1,2}/\d{2,4}
  • 纯年份：\d{2,4}
• 连接符：\s*(?:[-–]|to)\s*（匹配-、–、to三种连接方式，忽略前后空格）

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