14-索引扩展--数据分析

索引扩展

In [2]:

import pandas as pd

import numpy as np

In [5]:

s = pd.Series(np.arange(5),index = np.arange(5)[::-1],dtype='int64')

s

Out[5]:

4    0
3    1
2    2
1    3
0    4
dtype: int64

isin([1,3,4]) 查看数据表中是否有查看的数据

In [8]:

s.isin([1,3,4])

Out[8]:

4    False
3     True
2    False
1     True
0     True
dtype: bool

In [11]:

s[s.isin([1,3,4])]

Out[11]:

3    True
1    True
0    True
dtype: bool

MultiIndex.from_product([[0,1],['a','b','c']]) 二重索引

In [13]:

s2 = pd.Series(np.arange(6),index = pd.MultiIndex.from_product([[0,1],['a','b','c']]))

s2

Out[13]:

0  a    0
   b    1
   c    2
1  a    3
   b    4
   c    5
dtype: int32

In [14]:

s2.iloc[s2.index.isin([(1,'a'),(2,'b')])]

Out[14]:

1  a    3
dtype: int32

查看数据表中1a和0b两条数据

In [15]:

s2.iloc[s2.index.isin([(1,'a'),(0,'b')])]

Out[15]:

0  b    1
1  a    3
dtype: int32

In [18]:

s = pd.Series(np.arange(5),index = np.arange(5)[::-1],dtype='int64')

s

Out[18]:

4    0
3    1
2    2
1    3
0    4
dtype: int64

使用where帅选符合条件的数据

In [19]:

s[s>2]

Out[19]:

1    3
0    4
dtype: int64

In [20]:

dates = pd.date_range('20171022',periods=8)

In [22]:

df = pd.DataFrame(np.random.randn(8,4),index=dates,columns=['A','B','C','D'])

df

Out[22]:

	A	B	C	D
2017-10-22	-0.429368	-0.097481	0.643293	-0.984100
2017-10-23	0.223474	-0.897888	-0.240347	0.016184
2017-10-24	-0.961633	-1.063981	1.074710	0.360836
2017-10-25	1.173998	1.105386	-0.412135	-1.536320
2017-10-26	-0.656568	-0.938052	0.188153	-0.979891
2017-10-27	1.153825	-0.255656	1.194725	0.686401
2017-10-28	1.778536	1.809798	0.915557	-0.805165
2017-10-29	1.565753	0.705639	0.398115	1.356791

In [23]:

df.select(lambda x:x=='A',axis='columns')

d:\pythons\lib\site-packages\ipykernel_launcher.py:1: FutureWarning: 'select' is deprecated and will be removed in a future release. You can use .loc[labels.map(crit)] as a replacement
  """Entry point for launching an IPython kernel.

Out[23]:

	A
2017-10-22	-0.429368
2017-10-23	0.223474
2017-10-24	-0.961633
2017-10-25	1.173998
2017-10-26	-0.656568
2017-10-27	1.153825
2017-10-28	1.778536
2017-10-29	1.565753

• where() 在数据表中查看小于0的数据，大于0的数据默认显示NaN

In [24]:

df.where(df<0)

Out[24]:

	A	B	C	D
2017-10-22	-0.429368	-0.097481	NaN	-0.984100
2017-10-23	NaN	-0.897888	-0.240347	NaN
2017-10-24	-0.961633	-1.063981	NaN	NaN
2017-10-25	NaN	NaN	-0.412135	-1.536320
2017-10-26	-0.656568	-0.938052	NaN	-0.979891
2017-10-27	NaN	-0.255656	NaN	NaN
2017-10-28	NaN	NaN	NaN	-0.805165
2017-10-29	NaN	NaN	NaN	NaN

* where()  在数据表中将不符合条件的数据，默认显示为NaN替换成其他的数据。

In [26]:

df.where(df<0,'A')

Out[26]:

	A	B	C	D
2017-10-22	-0.429368	-0.0974806	A	-0.9841
2017-10-23	A	-0.897888	-0.240347	A
2017-10-24	-0.961633	-1.06398	A	A
2017-10-25	A	A	-0.412135	-1.53632
2017-10-26	-0.656568	-0.938052	A	-0.979891
2017-10-27	A	-0.255656	A	A
2017-10-28	A	A	A	-0.805165
2017-10-29	A	A	A	A

In [27]:

df = pd.DataFrame(np.random.rand(10,3),columns = list('abc'))

df

Out[27]:

	a	b	c
0	0.364628	0.665812	0.403687
1	0.142510	0.243596	0.809217
2	0.135084	0.884445	0.148034
3	0.129592	0.768869	0.163191
4	0.680117	0.216741	0.375793
5	0.324470	0.770352	0.522526
6	0.228491	0.985218	0.771736
7	0.813559	0.095543	0.009304
8	0.065755	0.262572	0.142525
9	0.603809	0.427960	0.414141

query() 查询数据

In [28]:

df.query('(a<b)')

Out[28]:

	a	b	c
0	0.364628	0.665812	0.403687
1	0.142510	0.243596	0.809217
2	0.135084	0.884445	0.148034
3	0.129592	0.768869	0.163191
5	0.324470	0.770352	0.522526
6	0.228491	0.985218	0.771736
8	0.065755	0.262572	0.142525

In [30]:

df.query('(a<b)&(b<c)')

Out[30]:

	a	b	c
1	0.14251	0.243596	0.809217

In [ ]:

​