python - Pandas 对行进行序列字符串匹配并获取最佳匹配行 ID

Question

假设我们有以下 pandas 数据框

import pandas as pd
data_dic = {
    "values": ['jk4', '293','814' ,'er b3', '1', " sas", '<', '37', '/',3, '5651 + sdfv 84083', '+', '814 gfj67 340f', "sas " ,'293', '<', 'df gfdh', ' .', ':9271', '1', '3-', '=', '5', '293', "sas "],
    "rowNr": [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24]
}
data = pd.DataFrame(data_dic)

给定一个特定的字符串，我们如何获得最匹配的行 ID？例如，假设输入字符串是:" sas 293 <"那么输出 ID 将为 [13,14,15]它与此数据框中的最佳匹配相对应。

示例2:对于输入字符串"814 gfj67 340f "输出将是 [12]

示例3:对于输入字符串". :92711"输出将是 [17,18,19]

Answer 1

好吧，这是我的尝试。

我仅通过计算匹配字符来计算最强的匹配，我继续进行所有可能的串联，并根据该分数选择最好的一个。

import pandas as pd
from itertools import product

data_dic = {
    "values": ['jk4', '293', '814', 'er b3', '1', " sas", '<', '37', '/', 3, '5651 + sdfv 84083', '+', '814 gfj67 340f',
               "sas ", '293', '<', 'df gfdh', ' .', ':9271', '1', '3-', '=', '5', '293', "sas "],
    "rowNr": [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24]
}
data = pd.DataFrame(data_dic)
data['values'] = data['values'].astype(str)

all_index_pairs = [(j, i) for i in range(len(data)) for j in range(i)]
all_concats = [''.join(data.loc[[*range(*pair)]]['values'].values) for pair in all_index_pairs]


def calc_match(s1, s2):
    return sum(1 for x, y in zip(s1.replace(' ', ''), s2.replace(' ', '')) if x == y)


def get_best_match(s):
    best_pair = max(zip(all_index_pairs, all_concats), key=lambda x: calc_match(s, x[1]))[0]
    return [*range(*best_pair)]


in1 = " sas 293 <"
in2 = "814 gfj67 340f "
in3 = ". :92711"

print(get_best_match(in1))
print(get_best_match(in2))
print(get_best_match(in3))

输出:

[13, 14, 15]
[12]
[17, 18, 19]