Question

如果我有数据框：

>>> import pandas as pd
>>> df = pd.DataFrame({'grp':['A', 'A', 'B', 'B', 'B'], 'pos' : [1, 2, 1, 2, 3], 'desc1' : ['X1', 'X2', 'Y1', 'Y2', 'Y3'], 'desc2' : ['A1', 'A2', 'A1', 'A2', 'A3']})
>>> df['desc'] = df.desc1 + ' (' + df.desc2 + ')'
>>> df = df.drop(columns=['desc1', 'desc2'])
>>> df
  grp  pos     desc
0   A    1  X1 (A1)
1   A    2  X2 (A2)
2   B    1  Y1 (A1)
3   B    2  Y2 (A2)
4   B    3  Y3 (A3)
>>>

我想将i转换为以下数据框：

  grp     pos1     pos2     pos3
0   A  X1 (A1)  X2 (A2)     None
1   B  Y1 (A1)  Y2 (A2)  Y3 (A3)

我想按“ grp”对所有内容进行分组，并希望每个位置在列单元格中都有desc。组的职位数量可变。怎么做？

致谢。

Answer 1

IIUC pd.crosstab

df1 = pd.crosstab(df.grp,df.pos,df.desc,aggfunc=lambda x : x)\
                .add_prefix('pos')\
                .reset_index()\
                .rename_axis(None,axis=1)

print(df1)
  grp     pos1     pos2     pos3
0   A  X1 (A1)  X2 (A2)      NaN
1   B  Y1 (A1)  Y2 (A2)  Y3 (A3)

Answer 2

或者您可以使用groupby和unstack解决它：

import pandas as pd
df = pd.DataFrame({'grp':['A', 'A', 'B', 'B', 'B'], 'pos' : [1, 2, 1, 2, 3], 'desc1' : ['X1', 'X2', 'Y1', 'Y2', 'Y3'], 'desc2' : ['A1', 'A2', 'A1', 'A2', 'A3']})
df['desc'] = df.desc1 + ' (' + df.desc2 + ')'
df = df.drop(columns=['desc1', 'desc2'])

df1 = df.groupby(['grp', 'pos'])['desc'].first().unstack('pos')
print(df1)
#Output:
pos        1        2        3
grp                           
A    X1 (A1)  X2 (A2)      NaN
B    Y1 (A1)  Y2 (A2)  Y3 (A3)

Python熊猫，如何转换数据框？

2 个答案: