Pandas——ix vs loc vs iloc

Different Choices for Indexing

loc——通过行标签索引行数据

• loc[1]表示索引的是第1行（index 是整数）

import pandas as pd
data = [[1,2,3],[4,5,6]]
index = [0,1]
columns=['a','b','c']
df = pd.DataFrame(data=data, index=index, columns=columns)
print df.loc[1]
'''
a    4
b    5
c    6
'''

• loc[‘d’]表示索引的是第’d’行（index 是字符）

import pandas as pd
data = [[1,2,3],[4,5,6]]
index = ['d','e']
columns=['a','b','c']
df = pd.DataFrame(data=data, index=index, columns=columns)
print df.loc['d']
'''
a    1
b    2
c    3
'''

• 如果想索引列数据，像这样做会报错

import pandas as pd
data = [[1,2,3],[4,5,6]]
index = ['d','e']
columns=['a','b','c']
df = pd.DataFrame(data=data, index=index, columns=columns)
print df.loc['a']
'''
KeyError: 'the label [a] is not in the [index]'
'''

• loc可以获取多行数据

import pandas as pd
data = [[1,2,3],[4,5,6]]
index = ['d','e']
columns=['a','b','c']
df = pd.DataFrame(data=data, index=index, columns=columns)
print df.loc['d':]
'''
   a  b  c
d  1  2  3
e  4  5  6
'''

• loc扩展——索引某行某列

import pandas as pd
data = [[1,2,3],[4,5,6]]
index = ['d','e']
columns=['a','b','c']
df = pd.DataFrame(data=data, index=index, columns=columns)
print df.loc['d',['b','c']]
'''
b    2
c    3
'''

• loc扩展——索引某列

import pandas as pd
data = [[1,2,3],[4,5,6]]
index = ['d','e']
columns=['a','b','c']
df = pd.DataFrame(data=data, index=index, columns=columns)
print df.loc[:,['c']]
'''
   c
d  3
e  6
'''

• 当然获取某列数据最直接的方式是df.[列标签]，但是当列标签未知时可以通过这种方式获取列数据。

• 需要注意的是，dataframe的索引[1:3]是包含1,2,3的，与平时的不同。

iloc——通过行号获取行数据

• 想要获取哪一行就输入该行数字

import pandas as pd
data = [[1,2,3],[4,5,6]]
index = ['d','e']
columns=['a','b','c']
df = pd.DataFrame(data=data, index=index, columns=columns)
print df.loc[1]
'''
a    4
b    5
c    6
'''

• 通过行标签索引会报错

import pandas as pd
data = [[1,2,3],[4,5,6]]
index = ['d','e']
columns=['a','b','c']
df = pd.DataFrame(data=data, index=index, columns=columns)
print df.iloc['a']
'''
TypeError: cannot do label indexing on <class 'pandas.core.index.Index'> with these indexers [a] of <type 'str'>
'''

• 同样通过行号可以索引多行

import pandas as pd
data = [[1,2,3],[4,5,6]]
index = ['d','e']
columns=['a','b','c']
df = pd.DataFrame(data=data, index=index, columns=columns)
print df.iloc[0:]
'''
   a  b  c
d  1  2  3
e  4  5  6
'''

• iloc索引列数据

import pandas as pd
data = [[1,2,3],[4,5,6]]
index = ['d','e']
columns=['a','b','c']
df = pd.DataFrame(data=data, index=index, columns=columns)
print df.iloc[:,[1]]
'''
   b
d  2
e  5
'''

ix——结合前两种的混合索引

• 通过行号索引

import pandas as pd
data = [[1,2,3],[4,5,6]]
index = ['d','e']
columns=['a','b','c']
df = pd.DataFrame(data=data, index=index, columns=columns)
print df.ix[1]
'''
a    4
b    5
c    6
'''

• 通过行标签索引

import pandas as pd
data = [[1,2,3],[4,5,6]]
index = ['d','e']
columns=['a','b','c']
df = pd.DataFrame(data=data, index=index, columns=columns)
print df.ix['e']
'''
a    4
b    5
c    6
'''