Series与DataFrame数据类型操作基础

Series创建方法：

#创建series
import pandas as pd
obj_list=[1,2,3,4]
obj_tuple=(4,5,6,7)
obj_dict={'a':[1,2],'b':[2,3],'c':[3,4],'d':[4,5]}
obj_series_list=pd.Series(obj_list)#通过列表创建series
obj_series_tuple=pd.Series(obj_tuple,index=list('abcd'))#通过元组创建series
obj_series_dict=pd.Series(obj_dict)#通过字典创建series
#定义属性
obj_series_list.index.name='zimu'
obj_series_list.name='data'

print "#通过列表创建series"
print obj_series_list
print "#通过元组创建series"
print obj_series_tuple
print "#通过字典创建series"
print obj_series_dict
#显示series类型及属性
print type(obj_series_list),obj_series_list.dtype
print obj_series_list.index,obj_series_list.index.name
print obj_series_list.values,obj_series_list.name

结果：

#通过列表创建series
zimu
0    1
1    2
2    3
3    4
Name: data, dtype: int64
#通过元组创建series
a    4
b    5
c    6
d    7
dtype: int32
#通过字典创建series
a    [1, 2]
b    [2, 3]
c    [3, 4]
d    [4, 5]
dtype: object
<class 'pandas.core.series.Series'> int64
RangeIndex(start=0, stop=4, step=1, name=u'zimu') zimu
[1 2 3 4] data

DataFrame创建方法

#创建dataframe
import pandas as pd
import numpy as np
obj_dict={'a':[1,2],'b':[2,3],'c':[3,4],'d':[4,5]}
obj_array=np.array([[1,2,3,4],[3,4,5,6]])
obj_series_1=pd.Series([11,12,13,14])
obj_series_2=pd.Series([21,22,23,24])
obj_dataframe_dict=pd.DataFrame(obj_dict)#通过字典创建dataframe
obj_dataframe_array=pd.DataFrame(obj_array,index=['one','two'])#通过矩阵创建dataframe
obj_dataframe_series=pd.DataFrame([obj_series_1,obj_series_2])#通过series创建dataframe
obj_dataframe_dataframe=pd.DataFrame(obj_dataframe_series,index=[0,1,'one'],columns=[0,1,2,'a'])#通过其他dataframe创建dataframe
print "#通过字典创建dataframe"
print obj_dataframe_dict
print "#通过矩阵创建dataframe"
print obj_dataframe_array
print "#通过series创建dataframe"
print obj_dataframe_series
print "#通过其他dataframe创建dataframe"
print obj_dataframe_dataframe
#dataframe属性
print obj_dataframe_dataframe.dtypes
print obj_dataframe_dataframe.values
print obj_dataframe_dataframe.columns
print obj_dataframe_dataframe.index

结果：

#通过字典创建dataframe
   a  b  c  d
0  1  2  3  4
1  2  3  4  5
#通过矩阵创建dataframe
     0  1  2  3
one  1  2  3  4
two  3  4  5  6
#通过series创建dataframe
    0   1   2   3
0  11  12  13  14
1  21  22  23  24
#通过其他dataframe创建dataframe
        0     1     2   a
0    11.0  12.0  13.0 NaN
1    21.0  22.0  23.0 NaN
one   NaN   NaN   NaN NaN
0    float64
1    float64
2    float64
a    float64
dtype: object
[[ 11.  12.  13.  nan]
 [ 21.  22.  23.  nan]
 [ nan  nan  nan  nan]]
Index([0, 1, 2, u'a'], dtype='object')
Index([0, 1, u'one'], dtype='object')

转自http://blog.csdn.net/ly_ysys629/article/details/54944153