基于多个标准的多列 Pandas 计数

Question

我有一个我想在下面计算的“单词”列表

word_list = ['one','two','three']

我在 pandas 数据框中有一列，下面有文字。

TEXT                                       | USER    | ID
-------------------------------------------|---------|------
"Perhaps she'll be the one for me."        | User 1  | 100
"Is it two or one?"                        | User 1  | 100
"Mayhaps it be three afterall..."          | User 2  | 150
"Three times and it's a charm."            | User 2  | 150
"One fish, two fish, red fish, blue fish." | User 2  | 150
"There's only one cat in the hat."         | User 3  | 200
"One does not simply code into pandas."    | User 3  | 200
"Two nights later..."                      | User 1  | 100
"Quoth the Raven... nevermore."            | User 2  | 150

我想要的输出如下所示，我想使用“TEXT”列中的数据计算与 word_list 中的任何单词相关的文本的唯一用户数。在计算了唯一用户之后，我还想计算与每条推文相关的关注者总和，与单词的唯一用户数相关联。

Word | Unique User Count | ID Sum
one  |      3            | 450
two  |      2            | 250
three|      1            | 150

有没有办法在 Python 2.7 中做到这一点？

Answer 1

我分解步骤

df.columns=['TEXT','USER','ID']

df[word_list]=df.TEXT.str.lower().apply(lambda x : pd.Series([x.find(y) for y in word_list])).ne(-1)
df1=df[['USER','one','two','three','ID']].set_index(['USER','ID']).astype(int).replace({0:np.nan})
Target=df1.stack().reset_index().groupby('level_2').agg({'USER':lambda x : len(set(x)),'ID':lambda x : sum(set(x))})
Target=Target.reset_index()
Target.columns=['Word','Unique User Count','ID Sum']
Target
Out[97]: 
    Word  Unique User Count  ID Sum
0    one                  3     450
1  three                  1     150
2    two                  2     250