pandas

Posted 2020-12-03 yinglai

pandas：数据分析

• pandas是一个强大的的python数据分析的工具包
• pandas是基于NumPy构建的
• pandas的主要功能：
  • 具备对其功能对数据结构DataFrame, Series
  • 集成时间序列功能
  • 提供丰富的数学运算和操作
  • 灵活处理缺失数据

安装方法:pip install pandas

引用方法：import pandas as pd

pandas:Series

Series是一种类似于一维数组的对象，有一组数据和一组与之相关的数据标签(索引)组成

Series比较像列表(数组)和字典的结合体

创建方式

In [62]: pd.Series([1,2,3,-6,7])
Out[62]: 
0    1
1    2
2    3
3   -6
4    7
dtype: int64


In [67]: pd.Series([1,2,3,-6,7],index=[‘a‘, ‘b‘, ‘c‘, ‘d‘, ‘e‘])
Out[67]: 
a    1
b    2
c    3
d   -6
e    7
dtype: int64


In [69]: pd.Series({‘a‘:11, ‘b‘:22})
Out[69]: 
a    11
b    22
dtype: int64

In [70]: pd.Series(0, index=[‘a‘, ‘b‘, ‘d‘])
Out[70]: 
a    0
b    0
d    0
dtype: int64

获取值数组和索引数组：values属性和index属性

In [71]: a = pd.Series([11,22,33,44,55], index=[‘a‘, ‘b‘, ‘c‘,‘d‘, ‘e‘])

In [72]: a
Out[72]: 
a    11
b    22
c    33
d    44
e    55
dtype: int64

In [73]: a.values
Out[73]: array([11, 22, 33, 44, 55])

In [74]: a.index
Out[74]: Index([u‘a‘, u‘b‘, u‘c‘, u‘d‘, u‘e‘], dtype=‘object‘) 

pandas:Series特性

Series支持NumPy模块的特性
从ndarray创建Series: Series(arr)
In [75]: a = pd.Series(np.array([11,22,33,44]))

In [76]: a
Out[76]: 
0    11
1    22
2    33
3    44
dtype: int64

与标量运算:sr*2
In [75]: a = pd.Series(np.array([11,22,33,44]))

In [76]: a
Out[76]: 
0    11
1    22
2    33
3    44
dtype: int64

In [77]: a+10
Out[77]: 
0    21
1    32
2    43
3    54
dtype: int64

In [79]: a*2
Out[79]: 
0    22
1    44
2    66
3    88
dtype: int64

两个Series运算:sr1+sr2
In [75]: a = pd.Series(np.array([11,22,33,44]))
In [80]: b = pd.Series([1,2,3,4])

In [81]: a+b
Out[81]: 
0    12
1    24
2    36
3    48
dtype: int64

索引:sr[0], sr[[0,2,4]]花式索引
In [75]: a = pd.Series(np.array([11,22,33,44]))
In [82]: a[0]
Out[82]: 11

In [85]: a[[0,2,3]]
Out[85]: 
0    11
2    33
3    44
dtype: int64

切片： sr[0:2]
In [75]: a = pd.Series(np.array([11,22,33,44]))
In [86]: a[0:3]
Out[86]: 
0    11
1    22
2    33
dtype: int64

布尔值过滤：sr[sr>0]
In [75]: a = pd.Series(np.array([11,22,33,44]))
In [87]: a[a>30]
Out[87]: 
2    33
3    44
dtype: int64

通用函数：np.abs(sr)
In [91]: a = pd.Series(-2)

In [92]: a
Out[92]: 
0   -2
dtype: int64

In [93]: np.abs(a)
Out[93]: 
0    2
dtype: int64

Series支持字典的特性

从字典创建Series: Series(dic)
In [94]: a = pd.Series({‘a‘:10, ‘b‘:20, ‘c‘:30})

In [95]: a
Out[95]: 
a    10
b    20
c    30
dtype: int64

in运算:‘a‘ in sr
In [96]: ‘c‘ in a
Out[96]: True

键索引:sr[‘a‘], sr[[‘a‘, ‘b‘, ‘c‘]]
In [97]: a[‘b‘]
Out[97]: 20

In [98]: a[[‘a‘, ‘c‘]]
Out[98]: 
a    10
c    30
dtype: int64

pandas:整数索引

整数索引的pandas对象错误示例
sr = np.Series(np.arange(4))
sr[-1]

如果索引是整数类型，则根据整数进行数据操作时总是面向标签的
loc 以标签解释
iloc 以下标解释

#loc以标签解释
In [123]: a = pd.Series([11,22,33,44], index=[11,22,33,44])
In [124]: a
Out[124]: 
11    11
22    22
33    33
44    44
dtype: int64

In [125]: a.loc[11]
Out[125]: 11
In [126]: a.loc[11:33]
Out[126]: 
11    11
22    22
33    33
dtype: int64

#iloc以下标解释
In [131]: a.iloc[-1]
Out[131]: 44

In [133]: a.iloc[0:3]
Out[133]: 
11    11
22    22
33    33
dtype: int64