pandas 的DataFrame学习

pandas 入门学习

Series 使用方法：

• Series定义
• Series 对象基本创建，值，索引如何访问
• Series 对象通过字典创建
• 判断Series对象是否存在缺失值

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

Pandas 的 DataFrame 学习

• DataFrame 定义：它是一个表格型的数据结构，它含有一组有序的列，每列可以是不同的值类型（布尔，字符串，数值等）
• DataFrame可以看作是由 Series组成的字典它们共用一个索引

DataFrame 的创建：最常用的方法是传入一个等长列表或者Numpy数组组成的 字典

data = {'state':['ohic','ohic','ohic','nevada','nevada'],
        'years':[2000,2001,2002,2003,2004],
        'pop':[[1.5, 1.7,3.5, 2.4,1.9]}

frame = DataFrame(data)

frame

	state	years	pop
0	ohic	2000	1.5
1	ohic	2001	1.7
2	ohic	2002	3.5
3	nevada	2003	2.4
4	nevada	2004	1.9

指定DataFrame列排列顺序

frame = DataFrame(data, columns = ['pop', 'state', 'years'])

frame

	pop	state	years
0	1.5	ohic	2000
1	1.7	ohic	2001
2	3.5	ohic	2002
3	2.4	nevada	2003
4	1.9	nevada	2004

fm = DataFrame(data, index = ['one','two', 'thress','four', 'five'], columns=['pop','state','years'])
# 指明行索引和列索引

fm

	pop	state	years
one	1.5	ohic	2000
two	1.7	ohic	2001
thress	3.5	ohic	2002
four	2.4	nevada	2003
five	1.9	nevada	2004

fm.columns

Index(['pop', 'state', 'years'], dtype='object')

fm.index

Index(['one', 'two', 'thress', 'four', 'five'], dtype='object')

# fm.index.name('a')

DataFrame 对于 列属性的处理：查看列的值，给某列赋值

fm['state'] # 通过列索引 查看某列值

one         ohic
two         ohic
thress      ohic
four      nevada
five      nevada
Name: state, dtype: object

fm.state # 通过列属性查看列的值

one         ohic
two         ohic
thress      ohic
four      nevada
five      nevada
Name: state, dtype: object

fm1 = DataFrame(data, index = ['one','two', 'thress','four', 'five'], columns=['pop','state','years','det'])

fm1

	pop	state	years	det
one	1.5	ohic	2000	NaN
two	1.7	ohic	2001	NaN
thress	3.5	ohic	2002	NaN
four	2.4	nevada	2003	NaN
five	1.9	nevada	2004	NaN

fm1.det = 19  # 通过属性给某列赋值 （思考：如果给指定行和列赋值呢？）

fm1

	pop	state	years	det
one	1.5	ohic	2000	19
two	1.7	ohic	2001	19
thress	3.5	ohic	2002	19
four	2.4	nevada	2003	19
five	1.9	nevada	2004	19

# 把列表或者数组赋值给某个列时，长度必须和DataFrame长度匹配，否则就会报错
v = [1,2,3,4,5]
fm1.det = v
fm1

	pop	state	years	det
one	1.5	ohic	2000	1
two	1.7	ohic	2001	2
thress	3.5	ohic	2002	3
four	2.4	nevada	2003	4
five	1.9	nevada	2004	5

# v = [1,2,3,5]#  Length of values does not match length of index
# fm1.det = v
# fm1

# 把Series对象付给某个列时，会精确匹配DataFrame的索引，所有空位会补上缺失值
# val = Series([1.2,1,3,2,4], index =['two','four','five'])
# # fm1.det = val
# fm1['det'] = val
# fm1

# 为了不存在的列赋值会创建新的列，del删除列
fm1['east'] = fm1.state=='ohic'
fm1

	pop	state	years	det	east
one	1.5	ohic	2000	1	True
two	1.7	ohic	2001	2	True
thress	3.5	ohic	2002	3	True
four	2.4	nevada	2003	4	False
five	1.9	nevada	2004	5	False

del fm1['east']

fm1

	pop	state	years	det
one	1.5	ohic	2000	1
two	1.7	ohic	2001	2
thress	3.5	ohic	2002	3
four	2.4	nevada	2003	4
five	1.9	nevada	2004	5

fm1.columns

Index(['pop', 'state', 'years', 'det'], dtype='object')

# 通过嵌套字典创建DataFrame,外层键为列名，内层键为行名
data1 = {'a':{'x':1, 'y':2,'z':3},
         'b':{'x':11,'y':12, 'z':13}}

fm2 = DataFrame(data1)

fm2

	a	b
x	1	11
y	2	12
z	3	13

DataFrame的名称属性：name, values,

fm1.index.name = 'Uid'
fm1.columns.name = 'co_name'

fm1

co_name	pop	state	years	det
Uid
one	1.5	ohic	2000	1
two	1.7	ohic	2001	2
thress	3.5	ohic	2002	3
four	2.4	nevada	2003	4
five	1.9	nevada	2004	5

fm1.values

array([[1.5, 'ohic', 2000, 1],
       [1.7, 'ohic', 2001, 2],
       [3.5, 'ohic', 2002, 3],
       [2.4, 'nevada', 2003, 4],
       [1.9, 'nevada', 2004, 5]], dtype=object)