Power BI Python 通过使用正则表达式解析现有列来创建新列

Question

我正在尝试通过使用正则表达式解析现有列来创建新列。 我想解析 key_set 列并创建 key1、key2、key3、key4 列并填充数据（如果该键存在于 key_set 中）。 某些密钥可能不存在或顺序可能不同。 这就是为什么我试图通过常规 exp 来实现这一点。 我必须使用如下所示的findall来完成它，因为我将在Power BI中使用这种方法。 任何帮助表示赞赏。

import re
import pandas as pd

d = {'name': ['John', 'Mark', 'Oz'],
     'key_set': ['https://test.com/direct/?key1=value344&key2=value566&key3=value577&key4=value888',
                 'https://test.com/direct/?key1=value113&key2=value332',
                 'https://test.com/direct/?key1=value111&key2=value222&key3=value333&key4=value444'] }

dataset = pd.DataFrame(data=d)
print(dataset)

keywords = ['key1', 'key2', 'key3', 'key4']
any_keyword = '|'.join(map(re.escape, keywords))
regex = "(" + any_keyword + ")=(.+?)(?=(?:" + any_keyword + ")=|$)"
# test_str='https://test.com/direct/?key1=value111&key2=value222&key3=value333&key4=value444'
# print([m.group().replace('&', '') for m in re.finditer(regex, test_str)])

print('')
print('====================================')
print('')

# APPROACH: This is how I test it and it works up to a point.
dataset['test1'] = dataset['key_set'].str.findall(regex)
print(dataset)

# I need to use the same approach and create one column per key attribute. How can I use above example and use it below.

# dataset['key1'] = ****
# dataset['key2'] = ****
# dataset['key3'] = ****
# dataset['key4'] = ****

Answer 1

def get_val( pattern, x ):
    """
    :param pattern: Regex pattern to match.
    :param x: String to search.
    :returns: Value if found, else None.
    """
    match = re.search( pattern, x )
    if match is None:
        return None
    
    return match.group( 2 )


for kn in range( 1, 5 ):
    key = f'key{kn}'  # key name

    # serach pattern ending with & or end of line
    pattern = f'({key})=(.*?)(?:&|$)'

    # apply get_val function over key_set for each key name
    df[ key ] = df[ 'key_set' ].apply( lambda x: get_val( pattern, x ) )

如果速度有问题，您还可以使用funtools.partial创建 function 以进行apply 。