將具有相同索引的列合並為 1 個單列（熊貓）

Question

我有一個dataframe我要改造。 我想將列合並在一起，並將具有相同 ID 的值顯示為單獨的行。

所以從這個：

id  children1       children2     children3
1   No Children    NaN           NaN
2   12-16 years    17+ years     Nan
3   No Children    Nan           Nan
4   5-7 years      8-11 years    12-16 years

為此：

id  children       
1   No Children    
2   12-16 years
2   17+ years     
3   No Children   
4   5-7 years
4   8-11 years
4   12-16 years

是否有捷徑可尋？

數據：

{'id': [1, 2, 3, 4],
 'children1': ['No Children', '12-16 years', 'No Children', '5-7 years'],
 'children2': [nan, '17+ years', nan, '8-11 years'],
 'children3': [nan, nan, nan, '12-16 years']}

Answer 1

我認為您正在尋找將列轉換為索引的stack方法。 現在要獲得確切的形狀，您需要set_index使用“id”設置索引（以便在stack ing 時將“id”保留為索引）。

out = df.set_index('id').stack().droplevel(1).rename('children').reset_index()

Output：

   id     children
0   1  No Children
1   2  12-16 years
2   2    17+ years
3   3  No Children
4   4    5-7 years
5   4   8-11 years
6   4  12-16 years

Answer 2

new = (df.set_index('id').agg(list,1)#Put all row values into a list except id
   .explode()#Ensure each element in a list is put in a row
   .replace('Nan', np.nan)# Make Nan -> NaN
   .dropna()#Drop all NaNs
   .to_frame('Children')#Rename column 0 to Childresn
  )

結果

     Children
id            
1   NoChildren
2   12-16years
2     17+years
3   NoChildren
4     5-7years
4    8-11years
4   12-16years
​