Python多行正则匹配交错结果：能否用单一模式匹配全部目标？

需求：匹配所有同时存在found、startprocess、endprocess记录的xxx.txt文件，示例文本如下：

texte="""
.....................................................................
...
... found:  a_dir/a_sub_dir/aaa.txt
.....................................................................
...
... found:  a_dir/a_sub_dir/bbb.txt
.....................................................................
...
... startprocess: aaa.txt
.....................................................................
...
... found:  a_dir/a_sub_dir/ccc.txt
.....................................................................
...
... startprocess: bbb.txt
.....................................................................
...
... endprocess: aaa.txt
.....................................................................
... endprocess: bbb.txt
"""

首次尝试：单一正则失效 #

使用以下正则尝试一次性匹配：

pattern=r' .*?(found:\s+.*?((?P<filename>[a-z]+\.txt)))(?:.*?)(?P<start>startprocess:\s+(?P=filename))(?:.*?)(?P<end>endprocess:\s+(?P=filename))'
for m in  re.finditer(pattern,texte, re.DOTALL):
    print(m.groupdict())

结果仅匹配到aaa.txt，遗漏了bbb.txt：

{'filename': 'aaa.txt', 'start': 'startprocess: aaa.txt', 'end': 'endprocess: aaa.txt'}

替代方案：循环匹配 #

先提取所有被发现的文件名，再逐个匹配对应流程：

list=[]
pattern_template=r'.*?(found:\s+.*?(?P<filename>@filename@))(?:.*?)(?P<start>startprocess:\s+(?P=filename))(?:.*?)(?P<end>endprocess:\s+(?P=filename))'
for filename in re.findall(r'(?:.*?)found:\s+.*?([a-z]+\.txt)',texte):
    pattern=pattern_template.replace('@filename@', filename)
    for m in re.finditer(pattern,texte, re.DOTALL):
        list.append(m.groupdict())
print(list)

结果成功匹配两个符合条件的文件：

[{'filename': 'aaa.txt', 'start': 'startprocess: aaa.txt', 'end': 'endprocess: aaa.txt'}, {'filename': 'bbb.txt', 'start': 'startprocess: bbb.txt', 'end': 'endprocess: bbb.txt'}]

核心疑问：是否存在单一正则模式，能一次性匹配所有符合条件的文件记录？

可以用单一正则实现一次性匹配，关键是通过负向预查限制匹配范围，避免跨文件的内容干扰。

可行正则写法 #

import re

texte="""
.....................................................................
...
... found:  a_dir/a_sub_dir/aaa.txt
.....................................................................
...
... found:  a_dir/a_sub_dir/bbb.txt
.....................................................................
...
... startprocess: aaa.txt
.....................................................................
...
... found:  a_dir/a_sub_dir/ccc.txt
.....................................................................
...
... startprocess: bbb.txt
.....................................................................
...
... endprocess: aaa.txt
.....................................................................
... endprocess: bbb.txt
"""

pattern = r'found:\s+.*?(?P<filename>[a-z]+\.txt)(?:(?!found:).)*?(?P<start>startprocess:\s+(?P=filename))(?:(?!found:).)*?(?P<end>endprocess:\s+(?P=filename))'
matches = re.finditer(pattern, texte, re.DOTALL)
result = [m.groupdict() for m in matches]
print(result)

运行结果 #

[{'filename': 'aaa.txt', 'start': 'startprocess: aaa.txt', 'end': 'endprocess: aaa.txt'}, {'filename': 'bbb.txt', 'start': 'startprocess: bbb.txt', 'end': 'endprocess: bbb.txt'}]

正则逻辑说明 #

• (?:(?!found:).)*?：这是负向预查结构，含义是"匹配任意字符，但只要下一个字符序列不是found:就继续"，确保每个匹配块只包含单个文件的完整流程记录，不会覆盖或跳过后续文件的内容。
• 首次尝试失败的原因：原正则中的.*?会从文本开头开始匹配，吃掉aaa.txt的所有相关内容后，剩余文本里bbb.txt的found记录无法满足正则从开头匹配的逻辑，导致遗漏。而负向预查限制匹配范围后，每个匹配都独立对应一个文件的流程，不会互相干扰。

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