python数据分析四：DataFrame基本结构

# -*- coding: utf-8 -*-
import pandas as pd

from pandas import Series,DataFrame

import numpy as np

'''
DataFrame是表格型的数据结构，它含有一组有序的列，每列是不同值类型
'''

#创建一个dataframed的表格
dict={'state':['a','a','a','b','b'],'year':[1998,1999,1998,1999,2000],'pop':[1.2,1.4,2.3,4.5,5.6]}
frame=DataFrame(dict)
print(frame)
'''
 state  year  pop
0     a  1998  1.2
1     a  1999  1.4
2     a  1998  2.3
3     b  1999  4.5
4     b  2000  5.6

'''
#指定列的顺序
data=DataFrame(dict,columns=['year','pop','state'])
print(data)
'''
 year  pop state
0  1998  1.2     a
1  1999  1.4     a
2  1998  2.3     a
3  1999  4.5     b
4  2000  5.6     b
'''



#传入多余的列，会产生NaN
data=DataFrame(dict,columns=['state','year','pop','debt'],index=['one','two','three','fore','five'])
print(data)
'''
   state  year  pop debt
one       a  1998  1.2  NaN
two       a  1999  1.4  NaN
three     a  1998  2.3  NaN
fore      b  1999  4.5  NaN
five      b  2000  5.6  NaN

'''
#获取表头
print(data.columns)#Index(['state', 'year', 'pop', 'debt'], dtype='object')

#获取列值
print(data['state'])
# one      a
# two      a
# three    a
# fore     b
# five     b

print(data.year)
# one      1998
# two      1999
# three    1998
# fore     1999
# five     2000

#获取行值
print(data.ix['one'])
# state       a
# year     1998
# pop       1.2
# debt      NaN

#给debt标量赋值
data['debt']=15.6
print(data)
#       state  year  pop  debt
# one       a  1998  1.2  15.6
# two       a  1999  1.4  15.6
# three     a  1998  2.3  15.6
# fore      b  1999  4.5  15.6
# five      b  2000  5.6  15.6

'''

给标量赋值索引

'''
data['debt']=np.arange(5)
print(data)
'''
 state  year  pop  debt
one       a  1998  1.2     0
two       a  1999  1.4     1
three     a  1998  2.3     2
fore      b  1999  4.5     3
five      b  2000  5.6     4

'''

'''
精准赋值给索引,没有赋值就是NaN
'''
val=Series([-1.22,22,33],index=['one','three','five'])
data['debt']=val
print(data)
'''
   state  year  pop   debt
one       a  1998  1.2  -1.22
two       a  1999  1.4    NaN
three     a  1998  2.3  22.00
fore      b  1999  4.5    NaN
five      b  2000  5.6  33.00

'''

'''
删除不符合条件的列
'''

data['deldata']=data.state=='a'
print(data)
'''
     state  year  pop   debt  deldata
one       a  1998  1.2  -1.22     True
two       a  1999  1.4    NaN     True
three     a  1998  2.3  22.00     True
fore      b  1999  4.5    NaN    False
five      b  2000  5.6  33.00    False
'''
del data['deldata']
print(data)
'''
      state  year  pop   debt
one       a  1998  1.2  -1.22
two       a  1999  1.4    NaN
three     a  1998  2.3  22.00
fore      b  1999  4.5    NaN
five      b  2000  5.6  33.00
'''


'''
另一种常见的数据形式,字典形式，行数据无值，以NaN显示

'''

dict={'hhb':{'2001':100,'2002':200},'zjx':{'2000':1.2,'2001':1.23,'2002':1.3}}
dict_data=DataFrame(dict)
print(dict_data)
#         hhb   zjx
# 2000    NaN  1.20
# 2001  100.0  1.23
# 2002  200.0  1.30
#转置
print(dict_data.T)
#      2000    2001   2002
# hhb   NaN  100.00  200.0
# zjx   1.2    1.23    1.3

'''
指定索引
'''
#错误
# dict_index=DataFrame(dict,index=[2001,2002,2003])
# print(dict_index)

#正确
pdata={'hhb':dict_data['hhb'][:-1],'zjx':dict_data['zjx'][:2]}
print(DataFrame(pdata))

#         hhb   zjx
# 2000    NaN  1.20
# 2001  100.0  1.23

'''
指定index和columns的属性
'''
dict_data.index.name='year'
dict_data.columns.name='name'
print(dict_data)
# name    hhb   zjx
# year
# 2000    NaN  1.20
# 2001  100.0  1.23
# 2002  200.0  1.30

#获取values值
print(dict_data.values)
# [[   nan   1.2 ]
#  [100.     1.23]
#  [200.     1.3 ]]

#data不同时，数据会兼容
print(data.values)
# [['a' 1998 1.2 -1.22]
#  ['a' 1999 1.4 nan]
#  ['a' 1998 2.3 22.0]
#  ['b' 1999 4.5 nan]
#  ['b' 2000 5.6 33.0]]