Python3.5 Pandas模块缺失值处理和层次索引实例详解

本文实例讲述了Python3.5 Pandas模块缺失值处理和层次索引。分享给大家供大家参考，具体如下：

1、pandas缺失值处理

import numpy as np
import pandas as pd
from pandas import Series,DataFrame

df3 = DataFrame([
 ["Tom",np.nan,456.67,"M"],
 ["Merry",34,345.56,np.nan],
 [np.nan,np.nan,np.nan,np.nan],
 ["John",23,np.nan,"M"],
 ["Joe",18,385.12,"F"]
],columns = ["name","age","salary","gender"])

print(df3)
print("=======判断NaN值=======")
print(df3.isnull())
print("=======判断非NaN值=======")
print(df3.notnull())
print("=======删除包含NaN值的行=======")
print(df3.dropna())
print("=======删除全部为NaN值的行=======")
print(df3.dropna(how="all"))

df3.ix[2,0] = "Gerry"    #修改第2行第0列的值
print(df3)

print("=======删除包含NaN值的列=======")
print(df3.dropna(axis=1))

运行结果：

   name   age  salary gender
0    Tom   NaN  456.67      M
1  Merry  34.0  345.56    NaN
2    NaN   NaN     NaN    NaN
3   John  23.0     NaN      M
4    Joe  18.0  385.12      F
=======判断NaN值=======
    name    age salary gender
0  False   True  False  False
1  False  False  False   True
2   True   True   True   True
3  False  False   True  False
4  False  False  False  False
=======判断非NaN值=======
    name    age salary gender
0   True  False   True   True
1   True   True   True  False
2  False  False  False  False
3   True   True  False   True
4   True   True   True   True
=======删除包含NaN值的行=======
  name   age  salary gender
4  Joe  18.0  385.12      F
=======删除全部为NaN值的行=======
    name   age  salary gender
0    Tom   NaN  456.67      M
1  Merry  34.0  345.56    NaN
3   John  23.0     NaN      M
4    Joe  18.0  385.12      F
    name   age  salary gender
0    Tom   NaN  456.67      M
1  Merry  34.0  345.56    NaN
2  Gerry   NaN     NaN    NaN
3   John  23.0     NaN      M
4    Joe  18.0  385.12      F
=======删除包含NaN值的列=======
    name
0    Tom
1  Merry
2  Gerry
3   John
4    Joe

import numpy as np
import pandas as pd
from pandas import Series,DataFrame

df4 = DataFrame(np.random.randn(7,3))
print(df4)

df4.ix[:4,1] = np.nan    #第0至3行，第1列的数据
df4.ix[:2,2] = np.nan
print(df4)

print(df4.fillna(0))    #将缺失值用传入的指定值0替换

print(df4.fillna({1:0.5,2:-1}))   #将缺失值按照指定形式填充

运行结果：

          0         1         2
0 -0.737618 -0.530302 -2.716457
1  0.810339  0.063028 -0.341343
2  0.070564  0.347308 -0.121137
3 -0.501875 -1.573071 -0.816077
4 -2.159196 -0.659185 -0.885185
5  0.175086 -0.954109 -0.758657
6  0.395744 -0.875943  0.950323
          0         1         2
0 -0.737618       NaN       NaN
1  0.810339       NaN       NaN
2  0.070564       NaN       NaN
3 -0.501875       NaN -0.816077
4 -2.159196       NaN -0.885185
5  0.175086 -0.954109 -0.758657
6  0.395744 -0.875943  0.950323
          0         1         2
0 -0.737618  0.000000  0.000000
1  0.810339  0.000000  0.000000
2  0.070564  0.000000  0.000000
3 -0.501875  0.000000 -0.816077
4 -2.159196  0.000000 -0.885185
5  0.175086 -0.954109 -0.758657
6  0.395744 -0.875943  0.950323
          0         1         2
0 -0.737618  0.500000 -1.000000
1  0.810339  0.500000 -1.000000
2  0.070564  0.500000 -1.000000
3 -0.501875  0.500000 -0.816077
4 -2.159196  0.500000 -0.885185
5  0.175086 -0.954109 -0.758657
6  0.395744 -0.875943  0.950323

2、pandas常用数学统计方法

import numpy as np
import pandas as pd
from pandas import Series,DataFrame

#pandas常用数学统计方法

arr = np.array([
 [98.5,89.5,88.5],
 [98.5,85.5,88],
 [70,85,60],
 [80,85,82]
])
df1 = DataFrame(arr,columns=["语文","数学","英语"])
print(df1)
print("=======针对列计算总统计值=======")
print(df1.describe())
print("=======默认计算各列非NaN值个数=======")
print(df1.count())
print("=======计算各行非NaN值个数=======")
print(df1.count(axis=1))

运行结果：

     语文    数学    英语
0  98.5  89.5  88.5
1  98.5  85.5  88.0
2  70.0  85.0  60.0
3  80.0  85.0  82.0
=======针对列计算总统计值=======
              语文         数学         英语
count   4.000000   4.000000   4.000000
mean   86.750000  86.250000  79.625000
std    14.168627   2.179449  13.412525
min    70.000000  85.000000  60.000000
25％    77.500000  85.000000  76.500000
50％    89.250000  85.250000  85.000000
75％    98.500000  86.500000  88.125000
max    98.500000  89.500000  88.500000
=======默认计算各列非NaN值个数=======
语文    4
数学    4
英语    4
dtype: int64
=======计算各行非NaN值个数=======
0    3
1    3
2    3
3    3
dtype: int64

import numpy as np
import pandas as pd
from pandas import Series,DataFrame、

#2.pandas相关系数与协方差
df2 = DataFrame({
 "GDP":[12,23,34,45,56],
 "air_temperature":[23,25,26,27,30],
 "year":["2001","2002","2003","2004","2005"]
})

print(df2)
print("=========相关系数========")
print(df2.corr())
print("=========协方差========")
print(df2.cov())
print("=========两个量之间的相关系数========")
print(df2["GDP"].corr(df2["air_temperature"]))
print("=========两个量之间协方差========")
print(df2["GDP"].cov(df2["air_temperature"]))

运行结果：

 GDP  air_temperature  year
0   12               23  2001
1   23               25  2002
2   34               26  2003
3   45               27  2004
4   56               30  2005
=========相关系数========
                      GDP  air_temperature
GDP              1.000000         0.977356
air_temperature  0.977356         1.000000
=========协方差========
                   GDP  air_temperature
GDP              302.5             44.0
air_temperature   44.0              6.7
=========两个量之间的相关系数========
0.97735555485
=========两个量之间协方差========
44.0

import numpy as np
import pandas as pd
from pandas import Series,DataFrame

#3.pandas唯一值、值计数及成员资格

df3 = DataFrame({
 "order_id":["1001","1002","1003","1004","1005"],
 "member_id":["m01","m01","m02","m01","m02",],
 "order_amt":[345,312.2,123,250.2,235]
})

print(df3)

print("=========去重后的数组=========")
print(df3["member_id"].unique())

print("=========值出现的频率=========")
print(df3["member_id"].value_counts())

print("=========成员资格=========")
df3 = df3["member_id"]
mask = df3.isin(["m01"])
print(mask)
print(df3[mask])

运行结果：

 member_id  order_amt order_id
0       m01      345.0     1001
1       m01      312.2     1002
2       m02      123.0     1003
3       m01      250.2     1004
4       m02      235.0     1005
=========去重后的数组=========
['m01' 'm02']
=========值出现的频率=========
m01    3
m02    2
Name: member_id, dtype: int64
=========成员资格=========
0     True
1     True
2    False
3     True
4    False
Name: member_id, dtype: bool
0    m01
1    m01
3    m01
Name: member_id, dtype: object

3、pandas层次索引

import numpy as np
import pandas as pd
from pandas import Series,DataFrame

#3.pandas层次索引
data = Series([998.4,6455,5432,9765,5432],
      index=[["2001","2001","2001","2002","2002"],
      ["苹果","香蕉","西瓜","苹果","西瓜"]]
      )
print(data)

df4 = DataFrame({
 "year":[2001,2001,2002,2002,2003],
 "fruit":["apple","banana","apple","banana","apple"],
 "production":[2345,5632,3245,6432,4532],
 "profits":[245.6,432.7,534.1,354,467.8]
})

print(df4)
print("=======层次化索引=======")
df4 = df4.set_index(["year","fruit"])
print(df4)
print("=======依照索引取值=======")
print(df4.ix[2002,"apple"])
print("=======依照层次化索引统计数据=======")
print(df4.sum(level="year"))
print(df4.mean(level="fruit"))
print(df4.min(level=["year","fruit"]))

运行结果：

2001  苹果     998.4
      香蕉    6455.0
      西瓜    5432.0
2002  苹果    9765.0
      西瓜    5432.0
dtype: float64
    fruit  production  profits  year
0   apple        2345    245.6  2001
1  banana        5632    432.7  2001
2   apple        3245    534.1  2002
3  banana        6432    354.0  2002
4   apple        4532    467.8  2003
=======层次化索引=======
             production  profits
year fruit
2001 apple         2345    245.6
     banana        5632    432.7
2002 apple         3245    534.1
     banana        6432    354.0
2003 apple         4532    467.8
=======依照索引取值=======
production    3245.0
profits        534.1
Name: (2002, apple), dtype: float64
=======依照层次化索引统计数据=======
      production  profits
year
2001        7977    678.3
2002        9677    888.1
2003        4532    467.8
        production     profits
fruit
apple         3374  415.833333
banana        6032  393.350000
             production  profits
year fruit
2001 apple         2345    245.6
     banana        5632    432.7
2002 apple         3245    534.1
     banana        6432    354.0
2003 apple         4532    467.8

希望本文所述对大家Python程序设计有所帮助。

来源：https://blog.csdn.net/loveliuzz/article/details/78498121