如何在Pandas DataFrame中基于指定短语提取文本中的温度数值

核心用Pandas内置的str.extract方法配合正则表达式实现，无需自定义循环，是Pandas场景下最简洁的写法：

1. 提取所有匹配的温度数值（不分最高/最低） #

如果不需要区分high near对应高温、low near对应低温，只需要提取所有匹配的温度值，直接用如下代码：

# 假设存储天气文本的列名为 weather，提取结果存入 temperature 列
df['temperature'] = df['weather'].str.extract(r'(?:high|low) near (\d+)', expand=False).astype('Int64')

2. 分开提取最高温和最低温 #

如果需要分别存储高温、低温的数值，用如下代码：

# 提取最高温
df['high_temp'] = df['weather'].str.extract(r'high near (\d+)', expand=False).astype('Int64')
# 提取最低温
df['low_temp'] = df['weather'].str.extract(r'low near (\d+)', expand=False).astype('Int64')

代码说明 #

• 正则表达式中(?:high|low)是非捕获组，只做匹配不提取，后面的(\d+)是捕获组，会提取high near或者low near后面连续的数字
• expand=False表示返回Series而非DataFrame，直接赋值给新列更方便
• 用Int64类型（注意首字母大写）是为了兼容没有匹配到数值的空值场景，避免数值被强制转成float类型

测试示例 #

你可以用如下测试代码验证效果：

import pandas as pd
# 构造测试数据
data = {
    'weather': [
        'Sunny, with a high near 82. Light and variable wind becoming northwest 5 to 7 mph in the afternoon.',
        'A 50 percent chance of showers.  Partly sunny, with a high near 61.',
        'Clear, with a low near 48. Northeast wind around 3 mph.',
        'Rainy, high near 59, low near 42.'
    ]
}
df = pd.DataFrame(data)
# 提取温度
df['high_temp'] = df['weather'].str.extract(r'high near (\d+)', expand=False).astype('Int64')
df['low_temp'] = df['weather'].str.extract(r'low near (\d+)', expand=False).astype('Int64')
print(df)

运行后输出结果如下：

weather  high_temp  low_temp
0  Sunny, with a high near 82. Light and variable...         82      <NA>
1  A 50 percent chance of showers.  Partly sunny,...         61      <NA>
2  Clear, with a low near 48. Northeast wind arou...       <NA>        48
3                 Rainy, high near 59, low near 42.         59        42

内容的提问来源于stack exchange，提问作者Guayule