Question

我正在尝试将字符串合并到称为df的数据帧中。如下：

s=['vic','tory','ban','ana']
df=pd.DataFrame(s,columns=['Tokens'])

请注意，我将仅将其用于另一种语言，而不是英语。

我想做的是合并df列中的行，并检查字典中组合的单词，如果该单词存在，那么它将被保存到另一个数据集中，并且df中的单词部分也将被删除。例如，我将df [0]和df [1]组合在一起，它就变成了“胜利”，然后在字典中寻找它并找到了它。然后将从df中删除“ vic”和“ tory”。我应该如何解决呢？任何帮助表示赞赏。

Answer 1

如果您有字符串列表，并且想要检查连续字符串的组合是否构成一个单词，则可以遍历字符串并检查可能的组合。为此，您可以使用内置的python手段：

LIMIT = 3  # max amount of strings to combine


def process_strings(strings, words):

    ans = list()

    stop = len(strings)
    current = 0
    # iterate over strings
    while current < stop:
        word = ''
        counter = 0
        # iterate over LIMIT strings starting from current string
        while True:
            # check boundary conditions
            if counter >= LIMIT or current + counter >= stop:
                current += 1
                break
            word += strings[current + counter]
            # word found among words
            if word in words:
                current += 1 + counter
                ans.append(word)
                # print('found word: {}'.format(word))
                break
            # word not found
            else:
                counter += 1

    return ans


words = {'victory', 'banana', 'python'}
strings = [
    'vic', 'tory',
    'mo', 'th', 'er',
    'ban', 'ana',
    'pyt', 'on',
    'vict', 'ory',
    'pyt', 'hon',
    'vi', 'ct', 'or', 'y',
    'ba', 'na', 'na']

words_found = process_strings(strings, words)
print('found words:\n{}'.format(words_found))

输出：

found words:
['victory', 'banana', 'victory', 'python', 'banana']

编辑

修改后的版本，适用于1）任意数量的字符串可以组合，2）诸如words = {'victory', 'victor'}，strings = ['vi', 'ct', 'or', 'y']之类的情况-可以找到两个词：

def process_strings(strings, words):

    MAXLEN = max(map(len, words))

    ans = list()

    stop = len(strings)
    current = 0
    # iterate over strings
    while current < stop:
        word = ''
        counter = 0
        # iterate over some amount of strings starting from current string
        while True:
            # check boundary conditions
            if len(word) > MAXLEN or current + counter >= stop:
                current += 1
                break
            word += strings[current + counter]
            # word found among words
            if word in words:
                ans.append(word)
            # there is no case `word not found`, exit only by boundary condition (length of the combined substrings)
            counter += 1

    return ans

合并熊猫列中的行中的字符串

1 个答案: