将字符串和整数列相乘

Question

我有一个具有这种结构的DataFrame [1]，我想将字符串和整数列相乘。

+----------------------+------------+-------------------------+-----------+--+
| url                  | date       | word                    | mentioned |  |
|----------------------+------------+-------------------------+-----------+--+
| newspaperarticle.com | 2018-12-22 | [canada,house,micheal]  | [2,2,1]   |  |
| articleUSA.com       | 2018-12-23 | [new york,murder,angry] | [2,3,1]   |  |
+----------------------+------------+-------------------------+-----------+-

我想要列名中的单词数乘以

+----------------------+------------+-------------------------+-------+---+--+
| url                  | date       | word                            |mentioned 
|----------------------+------------+-------------------------+-------+---+--+
| newspaperarticle.com | 2018-12-22 | [canada,canada,house,..]        |[2,2,1]     
| articleUSA.com       | 2018-12-23 | [new york,new york,murder,..]   |[2,3,1]   
+----------------------+------------+-------------------------+-------+---+--+

到目前为止，我所做的是用不起作用的乘法方法将列相乘。我还尝试了for循环，对单个元素建立索引并将它们相乘，但始终使错误字符串脱离索引。

Answer 1

您可以explode并使用series.repeat，将聚合作为级别= 0上的列表：

s = [df[i].explode() for i in ['word','mentioned']]
df['word'] = s[0].repeat(s[1]).groupby(level=0).agg(list)

---

print(df)

                    url        date  \
0  newspaperarticle.com  2018-12-22   
1        articleUSA.com  2018-12-23   

                                                word  mentioned  
0            [canada, canada, house, house, micheal]  [2, 2, 1]  
1  [new york, new york, murder, murder, murder, a...  [2, 3, 1]  

注意：这是假设word和mentioned列是一系列列表，而不是列表的字符串表示形式。