Python 跨多行查找所有内容的正则表达式

Question

过去一周我尝试解决这个问题，但没有取得任何进展。非常感谢大家的帮助。

我有 1000 个包含以下文本的文件：

,,,,,,,,,,,,,,,,
,,,,,,,,,,,,,,,,
Location:,,,ADDRESS_HERE_THAT I WANT
BUT IT CAN ALSO BE ACROSS,
MULTIPLE LINES, BUT NOT A SPECIFIC SET OF LINES,
AND IT ENDS AS ABRUPTLY,,,,,,,,,,,,,
,,,,,,,,,,,,,,,,

但少数文件也有这种方式

,,,,,,,,,,,,,,,,
Location:,,,ADDRESS,IS,IN,ONE,LINE,,,,,,,,,,,,,
,,,,,,,,,,,,,,,,

我需要在 Python.

中使用正则表达式提取大写地址

从技术上讲，它是一个由非常旧的系统导出的 CSV 文件。它实际上不能用作 CSV，因此我选择提取字符串，假设它是一个纯文本文件。

我当前的代码是这样的，但我已经尝试了很多其他组合，但没有找到有效的解决方案。

location = re.findall(r'^Location:,,,(.*),,,,,,,,,,,,,\n$|^Location:,,,(.*)[\n.*]{1,2,3,4,5,6},,,,,,,,,,,,,', CSV, flags=re.DOTALL | re.MULTILINE)

我离得还近吗？或者有更好的方法解决这个问题吗？

在此感谢您的帮助。

Answer 1

这里有一个思路：可以使用简单的循环来检测和提取多行位置数据

# Test data
TEXT=""",,,,,,,,,,,,,,,,
,,,,,,,,,,,,,,,,
Location:,,,ADDRESS_HERE_THAT I WANT
BUT IT CAN ALSO BE ACROSS,
MULTIPLE LINES, BUT NOT A SPECIFIC SET OF LINES,
AND IT ENDS AS ABRUPTLY,,,,,,,,,,,,,
,,,,,,,,,,,,,,,,
,,,,,,,,,,,,,,,,
Location:,,,ADDRESS,IS,IN,ONE,LINE,,,,,,,,,,,,,
,,,,,,,,,,,,,,,,
"""

in_location = False
tmp_location = None

def extract_location(l):
    global in_location
    global tmp_location
    if l.startswith("Location:"):
        in_location = True
        tmp_location = []
        # special case
        if l.endswith(',,,,,,,,,,,,,'):
            print(l[13:-13])
            in_location = False
        else:
            tmp_location.append(l[13:]) # Don't need 'Location:,,,'
    else:
        if in_location:
            tmp_location.append(l)
            if l.endswith(',,,,,,,,,,,,,'):
                # The end
                in_location = False
                res =  " ".join(tmp_location)
                print(res[0:-13])  # Remove trailing commas


def main():
    for line in TEXT.split("\n"):
        extract_location(line)


if __name__ == "__main__":
    main()

假设它被保存到一个名为 concept.py、

的文件中

$ python3 concept.py
DDRESS_HERE_THAT I WANT BUT IT CAN ALSO BE ACROSS, MULTIPLE LINES, BUT NOT A SPECIFIC SET OF LINES, AND IT ENDS AS ABRUPTLY
DDRESS,IS,IN,ONE,LINE

Answer 2

鉴于您提供的虚拟数据：

s = ''',,,,,,,,,,,,,,,,
,,,,,,,,,,,,,,,,
Location:,,,ADDRESS_HERE_THAT I WANT
BUT IT CAN ALSO BE ACROSS,
MULTIPLE LINES, BUT NOT A SPECIFIC SET OF LINES,
AND IT ENDS AS ABRUPTLY,,,,,,,,,,,,,
,,,,,,,,,,,,,,,,
,,,,,,,,,,,,,,,,
Location:,,,ADDRESS,IS,IN,ONE,LINE,,,,,,,,,,,,,
,,,,,,,,,,,,,,,,'''

您可以使用以下正则表达式：

matches = re.findall(r'Location:((?:[^,]*,){16})', s, flags=re.MULTILINE)

这是比赛的样子：

>>> print('\n\n'.join(matches))
,,,ADDRESS_HERE_THAT I WANT
BUT IT CAN ALSO BE ACROSS,
MULTIPLE LINES, BUT NOT A SPECIFIC SET OF LINES,
AND IT ENDS AS ABRUPTLY,,,,,,,,,,

,,,ADDRESS,IS,IN,ONE,LINE,,,,,,,,,

接下来要做什么取决于原始文件中逗号的含义。例如，您可能希望将它们替换为空格：

addrs = [match.replace(',', ' ').strip() for match in matches]

看起来像这样：

>>> print('\n\n'.join(addrs))
ADDRESS_HERE_THAT I WANT
BUT IT CAN ALSO BE ACROSS 
MULTIPLE LINES  BUT NOT A SPECIFIC SET OF LINES 
AND IT ENDS AS ABRUPTLY

ADDRESS IS IN ONE LINE

Python 跨多行查找所有内容的正则表达式

Python regex to findall across multiple lines

python

regex

csv

screen-scraping