将 pandas df 中的句子拆分为单词行并为每个句子编号

Question

我是 python 的新手，我有一个 pandas 数据框，如下所示：

df =
sn  sent                                                    ent
0   ['an', 'apple', 'is', 'an', 'example', 'of', 'what?']   ['O', 'F', '0', '0', '0', 'O', 'O']
1   ['a', 'potato', 'is', 'an', 'example', 'of', 'what?']   ['O', 'V', '0', '0', '0', 'O', 'O']

我想创建另一个 pandas 数据框，如下所示：

newdf= 
sn  sent        ent
0   an          O
    apple       F
    is          O
    an          O
    example     O
    of          O
    what?       O
1   a           O
    potato      V
    is          O
    an          O
    example     O
    of          O
    what?       O

我试过这段代码，结果如下所示

df.set_index('sn')
            .stack()
            .str.split(expand=True)
            .stack()
            .unstack(level=1)
            .reset_index(level=0, drop=0)

它接近我想要的，但似乎可以弄清楚其余部分

sn  sent        ent
0   ['an',      ['O',
0   'apple',    'F',
0   'is',       'O',
0   'an',       'O',
0   'example',  'O',
0   'of',       'O',
0   'what?',    'O',
1   'a',        'O',
1   'potato',   'V',
1   'is',       'O',
1   'an',       'O',
1   'example',  'O',
1   'of',       'O',
1   'what?']    'O']

非常感谢任何指点

Answer 1

df = pd.DataFrame({'sn': [0,1],
                  'sent': [['an', 'apple', 'is', 'an', 'example', 'of', 'what?'], ['a', 'potato', 'is', 'an', 'example', 'of', 'what?']],
                  'ent': [['O', 'F', '0', '0', '0', 'O', 'O'], ['O', 'V', '0', '0', '0', 'O', 'O']]})

df.apply(pd.Series.explode).set_index('sn')

结果：

       sent ent
sn             
0        an   O
0     apple   F
0        is   0
0        an   0
0   example   0
0        of   O
0     what?   O
1         a   O
1    potato   V
1        is   0
1        an   0
1   example   0
1        of   O
1     what?   O

将 pandas df 中的句子拆分为单词行并为每个句子编号

split sentences in pandas df into rows of words and number each sentence

split

python-3.x

pandas

unnest