Pandas 第二节 数据结构-DataFrame

 理论：

DataFrame数据结构：

• 类似多维数组/表格数据
• 每列数据可以是不同的类型
• 索引包括行索引(index)和列索引(label/column)

DataFrame的构建以及操作：

1.构建DataFrame

• 通过ndarray/列表
• 通过dict通

2.获取列数据（Series类型 ）

df_obj[label] 或 df_obj.label

3.增加列数据

df_obj[new_label] = data

4.删除列

• df_obj.drop(columns=[],inplace=False)，当inplace=True时，原数据不会改变，否则不变。
• del df_obj[col_idx]

 

 

实验：

第五课数据分析工具Pandas基础

第二节 数据结构-DataFrame

数据结构--DataFrame

In [1]:

 

import pandas as pd

import numpy as np

构建DataFrame

In [3]:

 

# 通过ndarray

array = np.random.randn(5,4)

print(array)

[[-0.12816457  0.27077416 -0.61990247  0.4035906 ]
 [-0.26341507  0.04203491  0.77618217  1.37930502]
 [ 1.48660347  0.2923378  -0.54919946 -0.30086526]
 [-0.53059414 -1.56234117 -0.33324783  0.41363335]
 [ 0.12839676 -0.96041499 -0.67103782 -1.01363347]]

In [4]:

 

df_obj = pd.DataFrame(array)

df_obj

Out[4]:

	0	1	2	3
0	-0.128165	0.270774	-0.619902	0.403591
1	-0.263415	0.042035	0.776182	1.379305
2	1.486603	0.292338	-0.549199	-0.300865
3	-0.530594	-1.562341	-0.333248	0.413633
4	0.128397	-0.960415	-0.671038	-1.013633

In [5]:

 

print(df_obj)

          0         1         2         3
0 -0.128165  0.270774 -0.619902  0.403591
1 -0.263415  0.042035  0.776182  1.379305
2  1.486603  0.292338 -0.549199 -0.300865
3 -0.530594 -1.562341 -0.333248  0.413633
4  0.128397 -0.960415 -0.671038 -1.013633

In [6]:

 

# 通过dict

dict_data = {'A': 1.,

             'B': pd.Timestamp('20200318'),

             'C': pd.Series(1,index=list(range(4)),dtype='float32'),

             'D': np.array([3] * 4,dtype='int32'),

             'E': ['Python','Java','C++','C#'],

             'F': 'ChinaHadoop'}

print(dict_data)

{'A': 1.0, 'B': Timestamp('2020-03-18 00:00:00'), 'C': 0    1.0
1    1.0
2    1.0
3    1.0
dtype: float32, 'D': array([3, 3, 3, 3]), 'E': ['Python', 'Java', 'C++', 'C#'], 'F': 'ChinaHadoop'}

In [7]:

 

df_obj2 = pd.DataFrame(dict_data)

df_obj2

Out[7]:

	A	B	C	D	E	F
0	1.0	2020-03-18	1.0	3	Python	ChinaHadoop
1	1.0	2020-03-18	1.0	3	Java	ChinaHadoop
2	1.0	2020-03-18	1.0	3	C++	ChinaHadoop
3	1.0	2020-03-18	1.0	3	C#	ChinaHadoop

获取行，列、值

In [9]:

 

df_obj2.columns

Out[9]:

Index(['A', 'B', 'C', 'D', 'E', 'F'], dtype='object')

In [10]:

 

df_obj2.index

Out[10]:

Int64Index([0, 1, 2, 3], dtype='int64')

In [11]:

 

df_obj2.values

Out[11]:

array([[1.0, Timestamp('2020-03-18 00:00:00'), 1.0, 3, 'Python',
        'ChinaHadoop'],
       [1.0, Timestamp('2020-03-18 00:00:00'), 1.0, 3, 'Java',
        'ChinaHadoop'],
       [1.0, Timestamp('2020-03-18 00:00:00'), 1.0, 3, 'C++',
        'ChinaHadoop'],
       [1.0, Timestamp('2020-03-18 00:00:00'), 1.0, 3, 'C#',
        'ChinaHadoop']], dtype=object)

预览数据

In [12]:

 

df_obj2.head(3)

Out[12]:

	A	B	C	D	E	F
0	1.0	2020-03-18	1.0	3	Python	ChinaHadoop
1	1.0	2020-03-18	1.0	3	Java	ChinaHadoop
2	1.0	2020-03-18	1.0	3	C++	ChinaHadoop

In [13]:

 

df_obj2.tail(3)

Out[13]:

	A	B	C	D	E	F
1	1.0	2020-03-18	1.0	3	Java	ChinaHadoop
2	1.0	2020-03-18	1.0	3	C++	ChinaHadoop
3	1.0	2020-03-18	1.0	3	C#	ChinaHadoop

In [14]:

 

df_obj2['E']

Out[14]:

0    Python
1      Java
2       C++
3        C#
Name: E, dtype: object

In [16]:

 

df_obj2.E

Out[16]:

0    Python
1      Java
2       C++
3        C#
Name: E, dtype: object

In [15]:

 

type(df_obj2['E'])

Out[15]:

pandas.core.series.Series

增加列数据

In [17]:

 

df_obj2['G'] = range(4)

In [18]:

 

df_obj2

Out[18]:

	A	B	C	D	E	F	G
0	1.0	2020-03-18	1.0	3	Python	ChinaHadoop	0
1	1.0	2020-03-18	1.0	3	Java	ChinaHadoop	1
2	1.0	2020-03-18	1.0	3	C++	ChinaHadoop	2
3	1.0	2020-03-18	1.0	3	C#	ChinaHadoop	3

删除列

In [21]:

 

df_obj3 = df_obj2.drop(columns=['B','G'])

In [22]:

 

df_obj3

Out[22]:

	A	C	D	E	F
0	1.0	1.0	3	Python	ChinaHadoop
1	1.0	1.0	3	Java	ChinaHadoop
2	1.0	1.0	3	C++	ChinaHadoop
3	1.0	1.0	3	C#	ChinaHadoop

In [23]:

 

df_obj2

Out[23]:

	A	B	C	D	E	F	G
0	1.0	2020-03-18	1.0	3	Python	ChinaHadoop	0
1	1.0	2020-03-18	1.0	3	Java	ChinaHadoop	1
2	1.0	2020-03-18	1.0	3	C++	ChinaHadoop	2
3	1.0	2020-03-18	1.0	3	C#	ChinaHadoop	3

In [ ]:

 

​