Question

我有一个使用numpy数组（MnthIdx，Val1，Val2，Val3）创建的数据框：

io.on('connection', socket =>{})

我是否有可能只用一步完成上述3个步骤。

Answer 1

您可以将pivot_table与3列用作参数values。 Aggfunc=[np.mean]可以省略，因为这是默认的聚合函数。最后，如果需要输出numpy array，请使用values：

print (pd.pivot_table(dfout3, index=['Idx'], values=['Col1', 'Col2', 'Col3']))

样品：

import pandas as pd
import numpy as np

MnthIdx = [1,2,2,3,3]
Val1 =    [2,5,2,3,4]
Val2 =    [6,1,5,3,5]
Val3 =    [3,9,5,7,8]


dfout3 = pd.DataFrame({'Idx': MnthIdx,
                       'Col1': Val1,
                       'Col2': Val2,
                       'Col3': Val3})

MeanTable1 = pd.pivot_table(dfout3, index=['Idx'], values=['Col1'], aggfunc=[np.mean])
MeanVal1 = np.asarray(MeanTable1['mean'])
MeanTable2 = pd.pivot_table(dfout3, index=['Idx'], values=['Col2'], aggfunc=[np.mean])
MeanVal2 = np.asarray(MeanTable2['mean'])
MeanTable3 = pd.pivot_table(dfout3, index=['Idx'], values=['Col3'], aggfunc=[np.mean])
MeanVal3 = np.asarray(MeanTable3['mean'])

print (MeanTable1)
    mean
    Col1
Idx     
1    2.0
2    3.5
3    3.5

print (MeanTable2)
    mean
    Col2
Idx     
1      6
2      3
3      4

print (MeanTable3)
    mean
    Col3
Idx     
1    3.0
2    7.0
3    7.5

print (pd.pivot_table(dfout3, index=['Idx'], values=['Col1', 'Col2', 'Col3']))
     Col1  Col2  Col3
Idx                  
1     2.0   6.0   3.0
2     3.5   3.0   7.0
3     3.5   4.0   7.5

print (pd.pivot_table(dfout3, index=['Idx'], values=['Col1', 'Col2', 'Col3']).values)
[[ 2.   6.   3. ]
 [ 3.5  3.   7. ]
 [ 3.5  4.   7.5]]

Answer 2

关闭jezael回答：

df    = pd.pivot_table(dfout3, index=['Idx'], values=['Col1', 'Col2', 'Col3'])
means = [ np.asarray(df[x]) for x in list(df)]
'MeanTable1','MeanTable2','MeanTable3' = means

或

(MeanTable1,MeanTable2,MeanTable3) = [ np.asarray(df[x]) for x in list(df)]

这将为您提供数组方式。

熊猫意味着跨越多个列

2 个答案: