用Python做数据分析之DataFrame3——数据选取

准备数据

import pandas as pd
import numpy as  np
df =pd.read_excel('filePath/test.xlsx')   #导入数据

print df

                   A         B         C
2017-01-01  0.743068  0.775753  0.586364
2017-01-02  0.726336  0.917315  0.770945
2017-01-03  0.448482  0.062748  0.792973
2017-01-04  0.481502  0.219382  0.835761
2017-01-05  0.475752  0.966919  0.491558
2017-01-06  0.885991  0.252072  0.913809
2017-01-07  0.076248  0.374731  0.595837
2017-01-08  0.395501  0.733482  0.228993
2017-01-09  0.390069  0.493331  0.069293
2017-01-10  0.679217  0.538165  0.376052

单列选取

#根据列名选取单列数据，并以Series的形式返回列数据
series_A =df['A']
print series_A

2017-01-01    0.743068
2017-01-02    0.726336
2017-01-03    0.448482
2017-01-04    0.481502
2017-01-05    0.475752
2017-01-06    0.885991
2017-01-07    0.076248
2017-01-08    0.395501
2017-01-09    0.390069
2017-01-10    0.679217
Freq: D, Name: A, dtype: float64

#根据列名选取单列数据，与上述方法的区别就是返回数据的类型是numpy.ndarray
array_A =df['A'].values   
print array_A,type(array_A)

[ 0.74306764  0.72633612  0.44848208  0.48150193  0.47575176  0.88599117
  0.07624825  0.39550126  0.39006884  0.67921732] <type 'numpy.ndarray'>

多列选取

#选取多列数据 ,返回是Dataframe
print  df[['A','B']]

                   A         B
2017-01-01  0.743068  0.775753
2017-01-02  0.726336  0.917315
2017-01-03  0.448482  0.062748
2017-01-04  0.481502  0.219382
2017-01-05  0.475752  0.966919
2017-01-06  0.885991  0.252072
2017-01-07  0.076248  0.374731
2017-01-08  0.395501  0.733482
2017-01-09  0.390069  0.493331
2017-01-10  0.679217  0.538165

单行选取

#根据行索引获取单行数据，返回是Series
series_row = df.loc[dates[2]]
print series_row,type(series_row)

A    0.448482
B    0.062748
C    0.792973
Name: 2017-01-03 00:00:00, dtype: float64 <class 'pandas.core.series.Series'>

#根据行数选取单行数据，返回是Series
series_row = df.iloc[3]  #返回第4行数据
print series_row,type(series_row)

A    0.481502
B    0.219382
C    0.835761
Name: 2017-01-04 00:00:00, dtype: float64 <class 'pandas.core.series.Series'>

#选取单行数据，返回是ndarray
array_row =df.iloc[3].values
print array_row,type(array_row)

[ 0.48150193  0.21938152  0.83576114] <type 'numpy.ndarray'>

单个元素选取

#通过行数列数获取某个具体位置的数据，用iloc[i,j]或者iat[i,j]
iloc00 =df.iloc[0,0]   #第1行第1列
print iloc00,type(iloc00)
iat11 =df.iat[1,1]     #第2行第2列
print iat11,type(iat11)

0.74306764384 <type 'numpy.float64'>
0.917314866922 <type 'numpy.float64'>

#通过行索引与列名获取数据，用at
print df.at[dates[2],'B']

0.0627479827402

print(df.loc['20170101':'20170103',['A','B']])

                   A         B
2017-01-01  0.743068  0.775753
2017-01-02  0.726336  0.917315
2017-01-03  0.448482  0.062748