如何在 Python 中修复我的重新编译语句

Question

I have a text file and I am using re to locate a specific section of text (a list containing water usage in different towns) and putting the information into a pandas dataframe. The text list is ordered using letters eg (a), (b), (c) etc. The code works fine and returns all the information I need into the dataframe up until the ordering switches to double letters eg (aa), (ab), (ac) etc.

How can I fix my re statement so that it also works for double lettered indexes in the text list?

Here is the code:

pattern = regex.compile('\d+ (?=ML\/year)|(?<= in the |the )[\w \/\(\)]+')
    columns = ('Water Usage', 'Town')

    res = [dict(zip(columns, pattern.findall(line))) for line in finalText.splitlines() if pattern.match(line)]
    df = pd.DataFrame(res)

    return df

And here is an example of the text:

(w) 218 ML/year in the Murrumbidgee I Water Source,
(x) 133 ML/year in the Murrumbidgee II Water Source,
(y) 116 ML/year in the Murrumbidgee III Water Source,
(z) 73 ML/year in the Murrumbidgee North Water Source,
(aa) 476 ML/year in the Murrumbidgee Western Water Source,
(ab) 92 ML/year in the Muttama Water Source,
(ac) 150 ML/year in the Numeralla East Water Source,

As I said, it works for all the rows with single letter indexes but doesn't for double letters.

Answer 1

You can use https://regex101.com/ or https://regexr.com/ to troubleshoot your regular expression. Here's one that matches the key components.

^\\([^)]+\\)\\s+(\\S+)\\s+(.*\\/year)\\s+in the\\s+(.*),

Answer 2

Python re module doesn't allow variable width pattern in look behind assertions.
correcting it, if you had used search() instead of match() it would have worked.

def create_df(finalText):
    pattern = re.compile('\d+ (?=ML\/year)|(?<= in the)[\w \/\(\)]+')
    columns = ('Water Usage', 'Town')
    res = [dict(zip(columns, pattern.findall(line))) for line in finalText.splitlines() if pattern.search(line)]
    df = pd.DataFrame(res)
    return df