Pandas库使用——各种索引

一、基本索引

import pandas as pd
import numpy as np

Series索引

ser_obj1 = pd.Series(range(5), index = ['a', 'b', 'c', 'd', 'e'])
print (ser_obj1.head())

a    0
b    1
c    2
d    3
e    4
dtype: int64

'''1. 行索引'''
print('行索引====================================================================')
print (ser_obj1['b'])
print (ser_obj1[0])
print('切片索引====================================================================')
'''2. 切片索引'''
print (ser_obj1[1:3])
print (ser_obj1['b':'d'])
print('不连续索引索引====================================================================')
#注意会不会包含尾巴。。
'''3. 不连续索引'''
print (ser_obj1[[0, 2, 4]])
print (ser_obj1[['a', 'e']])
print('布尔索引====================================================================')
'''4. 布尔索引'''
#是对里面的值进行判断，不是对索引
ser_bool = ser_obj1 > 2
print (ser_bool)
print (ser_obj1[ser_bool])#运用的方式
print (ser_obj1[ser_obj1 > 2])

行索引====================================================================
1
0
切片索引====================================================================
b    1
c    2
dtype: int64
b    1
c    2
d    3
dtype: int64
不连续索引索引====================================================================
a    0
c    2
e    4
dtype: int64
a    0
e    4
dtype: int64
布尔索引====================================================================
a    False
b    False
c    False
d     True
e     True
dtype: bool
d    3
e    4
dtype: int64
d    3
e    4
dtype: int64

Dataframe索引

df_obj1 = pd.DataFrame(np.random.randn(5,4), columns = ['a', 'b', 'c', 'd'])
print (df_obj1.head())

          a         b         c         d
0 -0.720611 -0.002543 -0.084255 -0.583719
1  0.174715  0.288090 -0.254462 -1.669668
2 -1.872654 -0.466842 -0.481040 -1.735558
3  1.147668  1.205686  0.625582 -2.111863
4  1.177882  0.049581 -0.404098 -0.497125

'''0. 转换成行索引的话===方式一：transpose或者T'''
bb=df_obj1.transpose()
print(df_obj1.T)
print(bb)
'''0. 转换成行索引的话===方式二：重构'''

          0         1         2         3         4
a -0.720611  0.174715 -1.872654  1.147668  1.177882
b -0.002543  0.288090 -0.466842  1.205686  0.049581
c -0.084255 -0.254462 -0.481040  0.625582 -0.404098
d -0.583719 -1.669668 -1.735558 -2.111863 -0.497125
          0         1         2         3         4
a -0.720611  0.174715 -1.872654  1.147668  1.177882
b -0.002543  0.288090 -0.466842  1.205686  0.049581
c -0.084255 -0.254462 -0.481040  0.625582 -0.404098
d -0.583719 -1.669668 -1.735558 -2.111863 -0.497125
'0. 转换成行索引的话===方式二：重构/重塑===还没有完全明白'

'''1. 列索引=====默认为列索引来操作'''
print ('列索引')
print(df_obj1.b)#当成属性来获取值
print('===================')
print (df_obj1['a']) # 返回Series类型
print (type(df_obj1['a']))
print (df_obj1[['a']]) # 返回DataFrame类型
print (type(df_obj1[['a']]))
'''2. 不连续索引'''
print ('不连续索引')
print (df_obj1[['a','c']])
print (df_obj1[['a','c']])

列索引
0   -0.002543
1    0.288090
2   -0.466842
3    1.205686
4    0.049581
Name: b, dtype: float64
===================
0   -0.720611
1    0.174715
2   -1.872654
3    1.147668
4    1.177882
Name: a, dtype: float64
<class 'pandas.core.series.Series'>
          a
0 -0.720611
1  0.174715
2 -1.872654
3  1.147668
4  1.177882
<class 'pandas.core.frame.DataFrame'>
不连续索引
          a         c
0 -0.720611 -0.084255
1  0.174715 -0.254462
2 -1.872654 -0.481040
3  1.147668  0.625582
4  1.177882 -0.404098
          a         c
0 -0.720611 -0.084255
1  0.174715 -0.254462
2 -1.872654 -0.481040
3  1.147668  0.625582
4  1.177882 -0.404098

三种索引方式

print(ser_obj1)
print('============================')
print(df_obj1)

a    0
b    1
c    2
d    3
e    4
dtype: int64
============================
          a         b         c         d
0 -0.720611 -0.002543 -0.084255 -0.583719
1  0.174715  0.288090 -0.254462 -1.669668
2 -1.872654 -0.466842 -0.481040 -1.735558
3  1.147668  1.205686  0.625582 -2.111863
4  1.177882  0.049581 -0.404098 -0.497125

'''1. 标签索引 loc===用的是index和column的值索引【标签索引】'''
# Series
print (ser_obj1['b':'d'])
print (ser_obj1.loc['b':'d'])
# DataFrame
print (df_obj1[['a']])
print (df_obj1.loc[0:2,'a':'c'])

b    1
c    2
d    3
dtype: int64
b    1
c    2
d    3
dtype: int64
          a
0 -0.720611
1  0.174715
2 -1.872654
3  1.147668
4  1.177882
          a         b         c
0 -0.720611 -0.002543 -0.084255
1  0.174715  0.288090 -0.254462
2 -1.872654 -0.466842 -0.481040

'''2. 位置索引 iloc====用的是index和column的默认0，1，2，3···的值索引【位置索引】'''
print (ser_obj1[1:3])
print (ser_obj1.iloc[1:3])
# DataFrame
print (df_obj1.iloc[0:2, 0:2])
# 注意和df_obj.loc[0:2, 'a']的区别 = = = 行和列：都不包括尾巴

b    1
c    2
dtype: int64
b    1
c    2
dtype: int64
          a         b
0 -0.720611 -0.002543
1  0.174715  0.288090

'''3. 混合索引 ix===先按标签索引loc尝试操作，然后再按位置索引iloc尝试操作'''
print (ser_obj1.ix[1:3])
print (ser_obj1.ix['b':'c'])
# DataFrame
print (df_obj1.ix[0:2, 0:3])

b    1
c    2
dtype: int64
b    1
c    2
dtype: int64
          a         b         c
0 -0.720611 -0.002543 -0.084255
1  0.174715  0.288090 -0.254462
2 -1.872654 -0.466842 -0.481040

二、Pandas层级索引

构造层级索引

ser_obj = pd.Series(np.random.randn(12),
                    index=[['a', 'a', 'a', 'b', 'b', 'b', 'c', 'c', 'c', 'd', 'd', 'd'],
                           [0, 1, 2, 0, 1, 2, 0, 1, 2, 0, 1, 2]])
#列表里面的列表
print (ser_obj)

a  0   -0.373441
   1    0.615976
   2    0.959092
b  0    1.743670
   1   -0.791517
   2   -0.774013
c  0    1.271094
   1   -0.723264
   2    0.253038
d  0   -0.767791
   1    0.419253
   2   -0.691644
dtype: float64

• MultiIndex索引对象

print (type(ser_obj.index))
print (ser_obj.index)

<class 'pandas.core.indexes.multi.MultiIndex'>
MultiIndex([('a', 0),
            ('a', 1),
            ('a', 2),
            ('b', 0),
            ('b', 1),
            ('b', 2),
            ('c', 0),
            ('c', 1),
            ('c', 2),
            ('d', 0),
            ('d', 1),
            ('d', 2)],
           )

选取子集

# 外层选取
print (ser_obj['c'])

0    1.271094
1   -0.723264
2    0.253038
dtype: float64

# 内层选取
print (ser_obj[:, 2])

a    0.959092
b   -0.774013
c    0.253038
d   -0.691644
dtype: float64

交换分层顺序

df_obj2 = pd.DataFrame(np.random.randn(6,4), columns = ['a', 'b', 'c', 'd'],index=[['a', 'a', 'a', 'b', 'b','b'],[1, 2, 3, 1, 2,3]])
#print (df_obj1.swaplevel())
print (df_obj2)

            a         b         c         d
a 1  0.273776 -0.018659  1.512727 -0.088091
  2 -0.410883 -1.488943  0.917268  1.179941
  3 -0.237532 -0.823717  0.189495  1.060476
b 1  0.727872  0.323352  0.443786  0.780510
  2 -1.407645 -0.059689  1.439843 -1.700740
  3 -0.377628 -0.137348 -0.739980  0.122528

交换并排序分层

print (df_obj2.swaplevel().sortvalues(by='a'))
#Series没有这个内置函数sortlevel===dataframe有

三、透视表和交叉表

dataset_path = './data/starcraft.csv'
df_data = pd.read_csv(dataset_path, usecols=['LeagueIndex', 'Age', 'HoursPerWeek', 
                                             'TotalHours', 'APM'])
print(df_data.head())

   LeagueIndex   Age  HoursPerWeek  TotalHours       APM
0            5  27.0          10.0      3000.0  143.7180
1            5  23.0          10.0      5000.0  129.2322
2            4  30.0          10.0       200.0   69.9612
3            3  19.0          20.0       400.0  107.6016
4            3  32.0          10.0       500.0  122.8908

透视表

pd.pivot_table(df_data, 
               index=['Age'],
               columns=['LeagueIndex'],
               aggfunc=np.sum)

28 rows × 21 columns

pd.pivot_table(df_data, 
               index=['Age'],
               columns=['LeagueIndex'],
               aggfunc=np.sum,
               fill_value=-100.)

28 rows × 21 columns

交叉表

pd.crosstab(df_data['LeagueIndex'],
            df_data['Age'])

7 rows × 28 columns

pd.crosstab(df_data['LeagueIndex'],
            [df_data['Age'], df_data['HoursPerWeek']])

7 rows × 325 columns