10分钟学会pandas库之一(查看和选择)

---

一、pandas是什么？

Pandas是一个开源，为Python (opens new window)编程语言提供高性能，易于使用的数据结构和数据分析工具。它的功能比EXCEL还要强大。

二、快速使用

1.引入库

代码如下：

import numpy as np
import pandas as pd
1
2

2.读取CSV文件数据

代码如下：

pd.read_csv(r'foo.csv')
          Unnamed: 0         A         B         C         D
0    ('2010-01-02',)  0.383275  0.694303  0.756312  0.325656
1    ('2010-01-03',)  0.789682  0.949437  0.018813  0.289073
2    ('2010-01-04',)  0.669891  0.575676  0.609218  0.371853
3    ('2010-01-05',)  0.335516  0.789114  0.201464  0.026223
4    ('2010-01-06',)  0.470739  0.080339  0.012590  0.636239
..               ...       ...       ...       ...       ...
995  ('2012-09-23',)  0.363259  0.533010  0.240098  0.171152
996  ('2012-09-24',)  0.238887  0.509116  0.847454  0.383001
997  ('2012-09-25',)  0.224259  0.645276  0.538157  0.539286
998  ('2012-09-26',)  0.556111  0.587099  0.599791  0.937187
999  ('2012-09-27',)  0.990479  0.073511  0.636951  0.485250
[1000 rows x 5 columns]
1
2
3
4
5
6
7
8
9
10
11
12
13
14

3.查看数据

head([n])查看前几行数据,默认n==5

df = pd.read_csv(r'foo.csv')
df.head()
        Unnamed: 0         A         B         C         D
0  ('2010-01-02',)  0.383275  0.694303  0.756312  0.325656
1  ('2010-01-03',)  0.789682  0.949437  0.018813  0.289073
2  ('2010-01-04',)  0.669891  0.575676  0.609218  0.371853
3  ('2010-01-05',)  0.335516  0.789114  0.201464  0.026223
4  ('2010-01-06',)  0.470739  0.080339  0.012590  0.636239
1
2
3
4
5
6
7
8

tail([n])查看末尾数据,也是默认n==5

df.tail()
          Unnamed: 0         A         B         C         D
995  ('2012-09-23',)  0.363259  0.533010  0.240098  0.171152
996  ('2012-09-24',)  0.238887  0.509116  0.847454  0.383001
997  ('2012-09-25',)  0.224259  0.645276  0.538157  0.539286
998  ('2012-09-26',)  0.556111  0.587099  0.599791  0.937187
999  ('2012-09-27',)  0.990479  0.073511  0.636951  0.485250
1
2
3
4
5
6
7

显示索引与列名：

df.index  
RangeIndex(start=0, stop=1000, step=1)  # 表示索引范围0到1000(左闭右开区间,步进step为1,即0，1，2，3...)

df.columns
Index(['Unnamed: 0', 'A', 'B', 'C', 'D'], dtype='object')
1
2
3
4
5

转置数据：

df.T
                        0                1    ...              998              999
Unnamed: 0  ('2010-01-02',)  ('2010-01-03',)  ...  ('2012-09-26',)  ('2012-09-27',)
A                  0.383275         0.789682  ...         0.556111         0.990479
B                  0.694303         0.949437  ...         0.587099         0.073511
C                  0.756312         0.018813  ...         0.599791         0.636951
D                  0.325656         0.289073  ...         0.937187          0.48525
[5 rows x 1000 columns]
1
2
3
4
5
6
7
8

按值排序：

按A列排序,默认升序

df.sort_values(by='A')
          Unnamed: 0         A         B         C         D
747  ('2012-01-19',)  0.000409  0.054803  0.720072  0.544639
992  ('2012-09-20',)  0.001385  0.442260  0.138257  0.960748
388  ('2011-01-25',)  0.002953  0.942049  0.832475  0.161548
395  ('2011-02-01',)  0.005425  0.753291  0.708701  0.950734
541  ('2011-06-27',)  0.008852  0.484906  0.535997  0.321985
..               ...       ...       ...       ...       ...
583  ('2011-08-08',)  0.995928  0.128598  0.819095  0.691981
191  ('2010-07-12',)  0.997268  0.463426  0.436351  0.983158
873  ('2012-05-24',)  0.998636  0.772841  0.217930  0.373149
246  ('2010-09-05',)  0.998801  0.305137  0.083263  0.520577
576  ('2011-08-01',)  0.999404  0.630375  0.796221  0.376956
[1000 rows x 5 columns]
1
2
3
4
5
6
7
8
9
10
11
12
13
14

按轴排序

二维数据行的方向代表0轴,列的方向代表1轴
下面表示按列的方向排序列索引0ABCD ->ABCD0

df.sort_index(axis=1)
            A         B         C         D       Unnamed: 0
0    0.383275  0.694303  0.756312  0.325656  ('2010-01-02',)
1    0.789682  0.949437  0.018813  0.289073  ('2010-01-03',)
2    0.669891  0.575676  0.609218  0.371853  ('2010-01-04',)
3    0.335516  0.789114  0.201464  0.026223  ('2010-01-05',)
4    0.470739  0.080339  0.012590  0.636239  ('2010-01-06',)
..        ...       ...       ...       ...              ...
995  0.363259  0.533010  0.240098  0.171152  ('2012-09-23',)
996  0.238887  0.509116  0.847454  0.383001  ('2012-09-24',)
997  0.224259  0.645276  0.538157  0.539286  ('2012-09-25',)
998  0.556111  0.587099  0.599791  0.937187  ('2012-09-26',)
999  0.990479  0.073511  0.636951  0.485250  ('2012-09-27',)
[1000 rows x 5 columns]
1
2
3
4
5
6
7
8
9
10
11
12
13
14

4.选择数据

按标签选择按

df.loc[0]获取标签为0的那行数据

df.loc[0]
Unnamed: 0    ('2010-01-02',)
A                    0.383275
B                    0.694303
C                    0.756312
D                    0.325656
Name: 0, dtype: object
1
2
3
4
5
6
7

选取多行多列
df.loc[:, [‘A’, ‘D’]] 选取AD两列的所有数据；

df.loc[:, ['A', 'D']]
            A         D
0    0.383275  0.325656
1    0.789682  0.289073
2    0.669891  0.371853
3    0.335516  0.026223
4    0.470739  0.636239
..        ...       ...
995  0.363259  0.171152
996  0.238887  0.383001
997  0.224259  0.539286
998  0.556111  0.937187
999  0.990479  0.485250
[1000 rows x 2 columns]
1
2
3
4
5
6
7
8
9
10
11
12
13
14

按位置选择

选择第三行的数据

df.iloc[2]
Unnamed: 0    ('2010-01-04',)
A                    0.669891
B                    0.575676
C                    0.609218
D                    0.371853
Name: 2, dtype: object
1
2
3
4
5
6
7

选取4，5行 第一列和第二列的数据

df.iloc[3:5, 0:2]
        Unnamed: 0         A
3  ('2010-01-05',)  0.335516
4  ('2010-01-06',)  0.470739
1
2
3
4

选取2，3，6行，第二三列的数据

df.iloc[3:5, 0:2]
        Unnamed: 0         A
3  ('2010-01-05',)  0.335516
4  ('2010-01-06',)  0.470739
df.iloc[[1,2,5], [1,3]]
          A         C
1  0.789682  0.018813
2  0.669891  0.609218
5  0.283397  0.085547
1
2
3
4
5
6
7
8
9

条件选择

A列大于0.7的所有行

df[df.A>0.7]
          Unnamed: 0         A         B         C         D
1    ('2010-01-03',)  0.789682  0.949437  0.018813  0.289073
6    ('2010-01-08',)  0.909243  0.089120  0.177339  0.580291
9    ('2010-01-11',)  0.866655  0.554461  0.498087  0.060520
13   ('2010-01-15',)  0.733417  0.372051  0.352662  0.134456
14   ('2010-01-16',)  0.973311  0.204662  0.786659  0.389995
..               ...       ...       ...       ...       ...
978  ('2012-09-06',)  0.855976  0.051711  0.289529  0.540159
981  ('2012-09-09',)  0.958141  0.729275  0.658488  0.079401
987  ('2012-09-15',)  0.939371  0.707939  0.068135  0.411151
994  ('2012-09-22',)  0.935422  0.334959  0.275858  0.652320
999  ('2012-09-27',)  0.990479  0.073511  0.636951  0.485250
[298 rows x 5 columns]
1
2
3
4
5
6
7
8
9
10
11
12
13
14

用isin()筛选

isin筛选E列里是aa或ee的数据

df2[df2['E'].isin(['aa', 'ee'])]
        Unnamed: 0         A         B         C         D   E
0  ('2010-01-02',)  0.383275  0.694303  0.756312  0.325656  aa
4  ('2010-01-06',)  0.470739  0.080339  0.012590  0.636239  ee
df2
        Unnamed: 0         A         B         C         D   E
0  ('2010-01-02',)  0.383275  0.694303  0.756312  0.325656  aa
1  ('2010-01-03',)  0.789682  0.949437  0.018813  0.289073  bb
2  ('2010-01-04',)  0.669891  0.575676  0.609218  0.371853  cc
3  ('2010-01-05',)  0.335516  0.789114  0.201464  0.026223  dd
4  ('2010-01-06',)  0.470739  0.080339  0.012590  0.636239  ee
5  ('2010-01-07',)  0.283397  0.927063  0.085547  0.608495  ff
df2[df2['E'].isin(['aa', 'ee'])]
        Unnamed: 0         A         B         C         D   E
0  ('2010-01-02',)  0.383275  0.694303  0.756312  0.325656  aa
4  ('2010-01-06',)  0.470739  0.080339  0.012590  0.636239  ee
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16

---

总结

以上就是今天要讲的内容，本文仅仅简单介绍了pandas的使用，未完待续系列。。。敬请期待！！！