Question

是否存在任何方法或函数来填充python中的数据透视表缺少的多列和多行？

import pandas as pd
import numpy as np
from io import StringIO

csvfile = StringIO("""Date;Cat;Type;Value
01-Jan;AA;S;1
02-Jan;AA;F;2
02-Jan;BB;T;3
04-Jan;BB;T;3
05-Jan;CC;T;2
05-Jan;DD;T;1
05-Jan;BB;S;4
05-Jan;AA;S;2
05-Jan;DD;S;4""")

df = pd.read_csv(csvfile, sep = ';')
pt = pd.pivot_table(df, values = 'Value', index=['Cat', 'Type'], columns= ['Date'], aggfunc = np.sum, fill_value = 0)
pt

上面的代码结果如下所示，对于某些Cat，缺少03-Jan列的Type列缺少值（F，S，T）：

Cat|Type|01-Jan|02-Jan|04-Jan|05-Jan|
---+----+------+------+------+------+
AA |F   |      |     2|      |      |
   |S   |     1|      |      |     2|
BB |S   |      |      |      |     4|
   |T   |      |     3|     3|      |
CC |T   |      |      |      |     2|
DD |S   |      |      |      |     4|
   |T   |      |      |      |     1|

但预期结果希望为：

Cat|Type|01-Jan|02-Jan|03-Jan|04-Jan|05-Jan|
---+----+------+------+------+------+------+
AA |F   |      |     2|      |      |      |
   |S   |     1|      |      |      |     2|
   |T   |      |      |      |      |      |
BB |F   |      |      |      |      |      |
   |S   |      |      |      |      |     4|
   |T   |      |     3|      |     3|      |
CC |F   |      |      |      |      |      |
   |S   |      |      |      |      |      |
   |T   |      |      |      |      |     2|
DD |F   |      |      |      |      |      |
   |S   |      |      |      |      |     4|
   |T   |      |      |      |      |     1|

Answer 1

pivot_table和Cat列中的值的所有组合都需要Type之后的reindex：

m = pd.MultiIndex.from_product([df['Cat'].unique(),df['Type'].unique()], names=pt.index.names)
pt = pt.reindex(m)
print (pt)
Date      01-Jan  02-Jan  04-Jan  05-Jan
Cat Type                                
AA  S        1.0     0.0     0.0     2.0
    F        0.0     2.0     0.0     0.0
    T        NaN     NaN     NaN     NaN
BB  S        0.0     0.0     0.0     4.0
    F        NaN     NaN     NaN     NaN
    T        0.0     3.0     3.0     0.0
CC  S        NaN     NaN     NaN     NaN
    F        NaN     NaN     NaN     NaN
    T        0.0     0.0     0.0     2.0
DD  S        0.0     0.0     0.0     4.0
    F        NaN     NaN     NaN     NaN
    T        0.0     0.0     0.0     1.0

Answer 2

只需先将df['Type']转换为Categorical：

df['Type'] = df['Type'].astype('category')

这迫使熊猫用pivot_table显示每个值。最好将Pandas将诸如'sum'之类的字符串转换为优化函数。这是一个演示：

df['Type'] = df['Type'].astype('category')

pt = pd.pivot_table(df, values='Value', index=['Cat', 'Type'],
                    columns='Date', aggfunc='sum', fill_value=0)

print(pt)

Date      01-Jan  02-Jan  04-Jan  05-Jan
Cat Type                                
AA  F          0       2       0       0
    S          1       0       0       2
    T          0       0       0       0
BB  F          0       0       0       0
    S          0       0       0       4
    T          0       3       3       0
CC  F          0       0       0       0
    S          0       0       0       0
    T          0       0       0       2
DD  F          0       0       0       0
    S          0       0       0       4
    T          0       0       0       1

如何在Python中为数据透视表填充缺少的多个列和行？

2 个答案: