Panda与numpy的不同: panda是基于Dataframe数据类型的，而numpy是基于ndarray的

import pandas as pd

1. 创建

pd.DataFrame(data=None,index=None,columns=None,dtype=None,copy=False)

pd.DataFrame([[1,2,3],[3,5,7]])

df = pd.DataFrame([[1,2,3],[3,5,7]],columns=['a','b','c'])

df

df.index

RangeIndex(start=0, stop=2, step=1)

type(df)

pandas.core.frame.DataFrame

df.index.values

array([0, 1], dtype=int64)

df.columns.values

array(['a', 'b', 'c'], dtype=object)

df = pd.read_csv('SalesJan2009.csv') #导入一个某超市营业表

2. 取值 赋值

df.iloc[1,:] #第一行的数据

Transaction_date                   2001/2/9 4:53
Product                                 Product1
Price                                       1200
Payment_Type                                Visa
Name                                      Betina
City                Parkville                   
State                                         MO
Country                            United States
Account_Created                    2001/2/9 4:42
Last_Login                         2001/2/9 7:49
Latitude                                  39.195
Longitude                               -94.6819
Name: 1, dtype: object

df.iloc[1,:]['Price'] #可以加列名 iloc=Integer Location

'1200'

df.loc[1:5,['Price']] #第一行的数据 注意loc 与iolc的区别

df.loc[:,'Transaction_date'] #或者直接 df.Transaction_date 或df['Transaction_date']

0        2001/2/9 6:17
1        2001/2/9 4:53
2       2001/2/9 13:08
3       2001/3/9 14:44
4       2001/4/9 12:56
5       2001/4/9 13:19
6       2001/4/9 20:11
7       2001/2/9 20:09
8       2001/4/9 13:17
9       2001/4/9 14:11
10       2001/5/9 2:42
11       2001/5/9 5:39
12       2001/2/9 9:16
13      2001/5/9 10:08
14      2001/2/9 14:18
15       2001/4/9 1:05
16      2001/5/9 11:37
17       2001/6/9 5:02
18       2001/6/9 7:45
19       2001/2/9 7:35
20      2001/6/9 12:56
21      2001/1/9 11:05
22       2001/5/9 4:10
23       2001/6/9 7:18
24       2001/2/9 1:11
25       2001/1/9 2:24
26       2001/7/9 8:08
27       2001/2/9 2:57
28      2001/1/9 20:21
29       2001/8/9 0:42
            ...       
968     2001/2/9 17:24
969     2001/7/9 18:15
970     2001/3/9 21:19
971      2001/4/9 5:13
972      2001/4/9 9:28
973      2001/2/9 4:34
974      1/22/09 11:10
975     2001/7/9 12:06
976      1/29/09 13:25
977       1/27/09 2:57
978     2001/7/9 13:19
979      1/23/09 10:04
980       1/19/09 4:55
981      1/28/09 22:02
982     2001/4/9 18:57
983    2001/12/9 20:31
984      1/24/09 12:00
985      1/28/09 11:19
986     2001/7/9 17:48
987      1/23/09 12:42
988     2001/7/9 19:48
989      1/26/09 11:19
990     2001/5/9 13:23
991      1/26/09 13:41
992      1/20/09 10:42
993      1/22/09 14:25
994       1/28/09 5:36
995      2001/1/9 4:24
996     2001/8/9 11:55
997    2001/12/9 21:30
Name: Transaction_date, Length: 998, dtype: object

df['Price2'] =  df.Price #为表格追加一列

df #出现了Price2一列

df.append(df)# 出现了index不唯一的错误

3. 复杂查询

# select * from df where Latitude > 30
df.loc[df.Latitude>30,:]

df.loc[(df.Latitude>30) & (df.Longitude >0) ,:].reset_index()

# 3.2 where Country like 'Aus%'
df.loc[df.Country.str.startswith('Aus')]