熊猫堆叠数据框

zac*_*ha2 2 python pandas

我有一个数据框,看起来像:

sensorId                   1      2      3   

9b:f3:55:19:00:4b:12:00    1      7      8    
bf:f3:55:19:00:4b:12:00    6      5      9   
da:f3:55:19:00:4b:12:00    1      1      2  
Run Code Online (Sandbox Code Playgroud)

我想用以下结构重构为数据框:

 sensorId                 y
 9b:f3:55:19:00:4b:12:00  1
 9b:f3:55:19:00:4b:12:00  7
 9b:f3:55:19:00:4b:12:00  8
 bf:f3:55:19:00:4b:12:00  6
 bf:f3:55:19:00:4b:12:00  5
 bf:f3:55:19:00:4b:12:00  9
 da:f3:55:19:00:4b:12:00  1
 da:f3:55:19:00:4b:12:00  1
 da:f3:55:19:00:4b:12:00  2
Run Code Online (Sandbox Code Playgroud)

我尝试使用df.stack(),但结果并不令人满意,因为它返回了pd.series:

9b:f3:55:19:00:4b:12:00  1      1
                         2      7
                         3      8
bf:f3:55:19:00:4b:12:00  1      6
                         2      5
                         3      9
da:f3:55:19:00:4b:12:00  1      1
                         2      1
                         3      2
Run Code Online (Sandbox Code Playgroud)

jez*_*ael 5

使用double Series.reset_index-first删除的第二级,MultiIndex然后使用第二转换SeriesDataFrame

df = df.stack().reset_index(level=1, drop=True).reset_index(name='y')
print (df)
                  sensorId  y
0  9b:f3:55:19:00:4b:12:00  1
1  9b:f3:55:19:00:4b:12:00  7
2  9b:f3:55:19:00:4b:12:00  8
3  bf:f3:55:19:00:4b:12:00  6
4  bf:f3:55:19:00:4b:12:00  5
5  bf:f3:55:19:00:4b:12:00  9
6  da:f3:55:19:00:4b:12:00  1
7  da:f3:55:19:00:4b:12:00  1
8  da:f3:55:19:00:4b:12:00  2
Run Code Online (Sandbox Code Playgroud)

编辑:

因为df.stackreturn MultiIndex,它的意思sensorId不是列,而是索引。

如果sensorId是列:

df = df.set_index('sensorId').stack().reset_index(level=1, drop=True).reset_index(name='y')
Run Code Online (Sandbox Code Playgroud)