如何在Pandas两列中替换多个复杂字符并解决re.error报错

需要在Pandas DataFrame的From和to列中，替换所有+符号以及列表S=['H2O','NADP(+)','NADPH','NAD(+)', 'NADH', 'H(+)']中的字符，但运行代码时出现re.error: nothing to repeat at position 2报错。

原始数据与错误代码 #

数据定义 #

import pandas as pd

classes = [('2.7.2.3', 'a primary alcohol',
            'an aldehyde'),
         ('2.7.1.3', 'a secondary alcohol',
          'a ketone'),
         ('3.1.1.3', 'an aldehyde + NADP(+)',
          'a 3-oxoacyl-[ACP] + NADPH'),
         ('3.1.1.3', '3-oxoacyl-[ACP] + NAD(+)',
          '2,3-dioxo-L-gulonate + NADH'),
         ('2.7.2.3', 'D-ribitol 5-phosphate + NADP(+)',
          'a primary alcohol + H(+)'),
         ('1.7.99.4', '2,3-dioxo-L-gulonate + NAD(+)',
          'D-ribulose 5-phosphate + NADH'),
         ('1.1.1.304', 'L-iditol + NAD(+)', ' H(+) + keto-L-sorbose + NADH'),
         ('2.7.4.3', 'H2O', 'oxidized coenzyme F420-1'),
         ('4.1.1.68', 'myo-inositol + NAD(+)', ' H(+) + NADH + a secondary alcohol')]
labels = ['Ko_EC','From', 'to']
alls = pd.DataFrame.from_records(classes, columns=labels)

错误代码 #

alls['From'] = alls['From'].str.replace("+", "")
alls['to'] = alls['to'].str.replace("+", "")

S = ['H2O','NADP()','NADPH','NAD()', 'NADH', 'H()']

alls

报错信息 #

re.error: nothing to repeat at position 2

期望结果 #

Ko_EC                           From                               to
0    2.7.2.3              a primary alcohol                      an aldehyde
1    2.7.1.3            a secondary alcohol                         a ketone
2    3.1.1.3                    an aldehyde                a 3-oxoacyl-[ACP]
3    3.1.1.3                3-oxoacyl-[ACP]             2,3-dioxo-L-gulonate 
4    2.7.2.3          D-ribitol 5-phosphate                a primary alcohol  
5   1.7.99.4           2,3-dioxo-L-gulonate           D-ribulose 5-phosphate
6  1.1.1.304                       L-iditol                   keto-L-sorbose  
7    2.7.4.3                                        oxidized coenzyme F420-1
8   4.1.1.68                   myo-inositol              a secondary alcohol

问题原因 #

1. +是正则表达式的特殊字符（表示匹配前序字符1次或多次），直接用str.replace会被当作正则语法解析，导致报错。
2. 原代码未正确处理目标列表中的字符，也没考虑+前后的空格，无法达到期望的替换效果。

解决方案 #

修正后代码 #

import pandas as pd
import re

# 数据定义（同原始数据）
classes = [('2.7.2.3', 'a primary alcohol',
            'an aldehyde'),
         ('2.7.1.3', 'a secondary alcohol',
          'a ketone'),
         ('3.1.1.3', 'an aldehyde + NADP(+)',
          'a 3-oxoacyl-[ACP] + NADPH'),
         ('3.1.1.3', '3-oxoacyl-[ACP] + NAD(+)',
          '2,3-dioxo-L-gulonate + NADH'),
         ('2.7.2.3', 'D-ribitol 5-phosphate + NADP(+)',
          'a primary alcohol + H(+)'),
         ('1.7.99.4', '2,3-dioxo-L-gulonate + NAD(+)',
          'D-ribulose 5-phosphate + NADH'),
         ('1.1.1.304', 'L-iditol + NAD(+)', ' H(+) + keto-L-sorbose + NADH'),
         ('2.7.4.3', 'H2O', 'oxidized coenzyme F420-1'),
         ('4.1.1.68', 'myo-inositol + NAD(+)', ' H(+) + NADH + a secondary alcohol')]
labels = ['Ko_EC','From', 'to']
alls = pd.DataFrame.from_records(classes, columns=labels)

# 定义需要替换的目标列表
targets = ['H2O','NADP(+)','NADPH','NAD(+)', 'NADH', 'H(+)']

# 生成正则匹配模式：匹配目标字符及前后空格，转义特殊字符避免正则错误
pattern = r'\s*(?:{})\s*'.format('|'.join([re.escape(t) for t in targets]))

# 批量处理From和to列
for col in ['From', 'to']:
    # 替换目标列表中的字符
    alls[col] = alls[col].str.replace(pattern, '', regex=True)
    # 替换剩余的+符号及前后空格
    alls[col] = alls[col].str.replace(r'\s*\+\s*', '', regex=True)
    # 清理首尾多余空格
    alls[col] = alls[col].str.strip()

print(alls)

代码说明 #

• re.escape(t)：转义目标字符中的特殊符号（如()），避免正则语法解析错误。
• r'\s*(?:{})\s*'：匹配目标字符前后的任意空格，非捕获组避免生成额外匹配内容。
• 先替换目标列表中的字符，再清理残留的+和空格，最后去除首尾空格，确保结果符合预期。

运行结果 #

Ko_EC                           From                               to
0    2.7.2.3              a primary alcohol                      an aldehyde
1    2.7.1.3            a secondary alcohol                         a ketone
2    3.1.1.3                    an aldehyde                a 3-oxoacyl-[ACP]
3    3.1.1.3                3-oxoacyl-[ACP]             2,3-dioxo-L-gulonate
4    2.7.2.3          D-ribitol 5-phosphate                a primary alcohol
5   1.7.99.4           2,3-dioxo-L-gulonate           D-ribulose 5-phosphate
6  1.1.1.304                       L-iditol                   keto-L-sorbose
7    2.7.4.3                                                   oxidized coenzyme F420-1
8   4.1.1.68                   myo-inositol              a secondary alcohol

内容的提问来源于stack exchange，提问作者yan wang