成為華爾街金融巨鱷第三課：

Pandas2:學會使用Pandas-DataFrame

import pandas as pd import numpy as np

一、DataFrame簡介和創(chuàng)建：二維數(shù)據(jù)對象

可以簡單理解為excel表格

創(chuàng)建方法一：利用字典創(chuàng)建

pd.DataFrame({"one":[1,2,3],'two':[4,5,6]}) onetwo012

1	4
2	5
3	6

和Series類似，我們可以為行指定索引

pd.DataFrame({"one":[1,2,3],'two':[4,5,6]},index=['a','b','c']) onetwoabc

1	4
2	5
3	6

創(chuàng)建方法二：利用Series創(chuàng)建

pd.DataFrame({'one':pd.Series([1,2,3],index=['a','b','c']),'two':pd.Series([1,2,3,4],index=['d','c','a','b'])}) onetwoabcd

1.0	3
2.0	4
3.0	2
NaN	1

可見DataFrame在創(chuàng)建時會自動進行索引對其

創(chuàng)建方法三：利用csv創(chuàng)建

pd.read_csv('test.csv') abc012

1	2	3
2	4	6
3	6	9

保存到csv

df = pd.DataFrame({'one':pd.Series([1,2,3],index=['a','b','c']),'two':pd.Series([1,2,3,4],index=['d','c','a','b'])}) df onetwoabcd

1.0	3
2.0	4
3.0	2
NaN	1

# 保存到csv df.to_csv('test2.csv')

二、DataFrame常見屬性

1、index、columns和vlues屬性

作用：index用來獲取 # 行 # 索引 ; columns獲取# 列 # 屬性；values用來獲取值 # 數(shù)組

df = pd.DataFrame({'one':pd.Series([1,2,3],index=['a','b','c']),'two':pd.Series([1,2,3,4],index=['d','c','a','b'])}) df onetwoabcd

1.0	3
2.0	4
3.0	2
NaN	1

# index屬性 df.index Index(['a', 'b', 'c', 'd'], dtype='object') # columns屬性 df.columns Index(['one', 'two'], dtype='object') # values屬性 df.values array([[ 1., 3.],[ 2., 4.],[ 3., 2.],[nan, 1.]])

2.T屬性

作用：轉(zhuǎn)置

df.T abcdonetwo

1.0	2.0	3.0	NaN
3.0	4.0	2.0	1.0

3.describe()方法

作用：返回詳細信息

df.describe() onetwocountmeanstdmin25%50%75%max

3.0	4.000000
2.0	2.500000
1.0	1.290994
1.0	1.000000
1.5	1.750000
2.0	2.500000
2.5	3.250000
3.0	4.000000

count: 該列數(shù)據(jù)共有多少條 mean：該列數(shù)據(jù)平均值 std：該列數(shù)據(jù)標準差 min：該列數(shù)據(jù)最小值 25%：該列數(shù)據(jù)從小到大25%位置上的數(shù) 50%：該列數(shù)據(jù)中位數(shù) 75%：該列數(shù)據(jù)從小到大75%位置上的數(shù) max：該列數(shù)據(jù)最大值

三、DataFrame索引和切片

df = pd.DataFrame({'one':pd.Series([1,2,3],index=['a','b','c']),'two':pd.Series([1,2,3,4],index=['d','c','a','b'])}) onetwoabcd

1.0	3
2.0	4
3.0	2
NaN	1

df取值可以采用df[x][y]的方法取值，表示取x列的y行，注意對比numpy，這里前一個中括號內(nèi)表示列

eg：取第one行第a列的1.0

df['one']['a'] 1.0

雖然這樣可以輕松地取到想要的值，但是一般情況下，我們不采取這種方式，因為會出現(xiàn)類似于Series整數(shù)索引的問題

于是，一般情況下我們還是使用loc和iloc來進行取值

用loc方式取值，在這種情況下逗號前表示行，逗號后表示列，和numpy類似

df.loc['a','one'] 1.0

！特別注意：DataFrame事實上是由n個Series對象所組成的，因此可以通過列索引直接取到某一列，卻不能通過行索引直接取到某一行,想要獲取某一行可以采取切片的方式

# 通過列名直接獲取某一列 df['one'] a 1.0 b 2.0 c 3.0 d NaN Name: one, dtype: float64 # 嚴禁通過行索引獲取某一行 df['a'] ---------------------------------------------------------------------------KeyError Traceback (most recent call last)c:\users\lenovo\appdata\local\programs\python\python37\lib\site-packages\pandas\core\indexes\base.py in get_loc(self, key, method, tolerance)2894 try: -> 2895 return self._engine.get_loc(casted_key)2896 except KeyError as err:pandas\_libs\index.pyx in pandas._libs.index.IndexEngine.get_loc()pandas\_libs\index.pyx in pandas._libs.index.IndexEngine.get_loc()pandas\_libs\hashtable_class_helper.pxi in pandas._libs.hashtable.PyObjectHashTable.get_item()pandas\_libs\hashtable_class_helper.pxi in pandas._libs.hashtable.PyObjectHashTable.get_item()KeyError: 'a'The above exception was the direct cause of the following exception:KeyError Traceback (most recent call last)<ipython-input-34-9637ce7feee6> in <module>1 # 嚴禁通過行索引獲取某一行 ----> 2 df['a']c:\users\lenovo\appdata\local\programs\python\python37\lib\site-packages\pandas\core\frame.py in __getitem__(self, key)2904 if self.columns.nlevels > 1:2905 return self._getitem_multilevel(key) -> 2906 indexer = self.columns.get_loc(key)2907 if is_integer(indexer):2908 indexer = [indexer]c:\users\lenovo\appdata\local\programs\python\python37\lib\site-packages\pandas\core\indexes\base.py in get_loc(self, key, method, tolerance)2895 return self._engine.get_loc(casted_key)2896 except KeyError as err: -> 2897 raise KeyError(key) from err2898 2899 if tolerance is not None:KeyError: 'a' # 通過切片獲取某一行 df.loc['a',:] one 1.0 two 3.0 Name: a, dtype: float64

根據(jù)我的嘗試，切片獲取：可以省略寫成df.loc[‘a(chǎn)’,]，甚至可以省略逗號df.loc[‘a(chǎn)’]，但是這與上面的注意違背所以盡量不要省略逗號

df = pd.DataFrame({'one':pd.Series([1,2,3],index=['a','b','c']),'two':pd.Series([1,2,3,4],index=['d','c','a','b'])}) df onetwoabcd

1.0	3
2.0	4
3.0	2
NaN	1

行/列索引部分除了常規(guī)索引，還可以是切片、布爾值索引、花式索引任意搭配

eg：

# 花式索引和切片搭配 df.loc[['a','b'],'one':'two'] onetwoab

1.0	3
2.0	4

注意點回顧：和Series一樣，在使用鍵索引時，是左閉右也閉的區(qū)間

四、數(shù)據(jù)對齊與缺失值處理

df = pd.DataFrame({'one':pd.Series([1,2,3],index=['a','b','c']),'two':pd.Series([1,2,3,4],index=['d','c','a','b'])}) df onetwoabcd

1.0	3
2.0	4
3.0	2
NaN	1

df2 = pd.DataFrame({'two':[1,2,3,4],'one':[5,6,7,8]},index = ['d','c','a','b']) df2 twoonedcab

1	5
2	6
3	7
4	8

df + df2 onetwoabcd

8.0	6
10.0	8
9.0	4
NaN	2

DataFrame遵循數(shù)據(jù)對齊的原則，在運算時，會行與列都分別對齊

df onetwoabcd

1.0	3
2.0	4
3.0	2
NaN	1

缺失值處理方法1：缺失值填充

# fillna(x)為缺失值填入x df.fillna(0) onetwoabcd

1.0	3
2.0	4
3.0	2
0.0	1

缺失值處理方法2：缺失值刪除

1.dropna()方法

# dropna()刪除存在缺失值所在行 df.dropna() onetwoabc

1.0	3
2.0	4
3.0	2

2.dropna()方法的how參數(shù)

df onetwoabcd

1.0	3
2.0	4
3.0	2
NaN	1

df.loc['c','two'] = np.nan df.loc['d','two'] = np.nan df onetwoabcd

1.0	3.0
2.0	4.0
3.0	NaN
NaN	NaN

# how=‘a(chǎn)ll’規(guī)定了只有該行全部為空才刪除，默認how的參數(shù)為any，即只要存在缺失值就刪除該行 df.dropna(how = 'all') onetwoabc

1.0	3.0
2.0	4.0
3.0	NaN

3.dropna()方法的axis參數(shù)

axis=0表示以行為單位，axis=1表示以列為單位# df.dropna(axis=1)表示將存在缺失值的列刪除，默認axis=0

df2 twoonedcab

1	5
2	6
3	7
4	8

df2.loc['a','two'] = np.nan df2 twoonedcab

1.0	5
2.0	6
NaN	7
4.0	8

# df.dropna(axis=1)表示將存在缺失值的列刪除，默認axis=0 df2.dropna(axis=1) onedcab

5

6

7

8

#另外，df中一樣提供了isnull（）和notnull（）使用方法和Series完全一致 df2.isnull() twoonedcab

False	False
False	False
True	False
False	False

五、DataFrame常見函數(shù)

1.求平均值

df = df2 df twoonedcab

1.0	5
2.0	6
NaN	7
4.0	8

df.mean() two 2.333333 one 6.500000 dtype: float64

mean():mean方法將忽略缺失值，計算出每一列的平均值，并返回一個Series對象

axis參數(shù)：axis=1可以按行求平均值 默認為0按列求平均值

df.mean(axis=1) d 3.0 c 4.0 a 7.0 b 6.0 dtype: float64

另外，sum，std等方法和mean()類似

2.按值排序 ： sort_values（）

df twoonedcab

1.0	5
2.0	6
NaN	7
4.0	8

# 按two這一列的值排序 df.sort_values(by = 'two') twoonedcba

1.0	5
2.0	6
4.0	8
NaN	7

# 按two這一列的值降序排序 df.sort_values(by = 'two',ascending=False) twoonebcda

4.0	8
2.0	6
1.0	5
NaN	7

# 按d這一行降序排序 df.sort_values(by = 'd',ascending=False,axis=1) onetwodcab

5	1.0
6	2.0
7	NaN
8	4.0

總結(jié)：

1.df.sort_values(）用以按值排序

2.by參數(shù)：指定按哪一列（行）排序

3.ascending參數(shù)：指定升序或降序排序，True為升序，False為降序，默認為True

4.axis參數(shù)：指定按行/列排序，axis=0按列排序，axis=1按行排序，默認為0，如果by的參數(shù)為行標簽，則必須賦值axis=1

5.關(guān)于缺失值NaN：如果存在缺失值，則缺失值不參與排序，統(tǒng)一放在最后面

3.按列排序 ： sort_index（）

df twoonedcab

1.0	5
2.0	6
NaN	7
4.0	8

df.sort_index() twooneabcd

NaN	7
4.0	8
2.0	6
1.0	5

# 降序排序 df.sort_index(ascending = False) twoonedcba

1.0	5
2.0	6
4.0	8
NaN	7

# 按one two列的順序排序 df.sort_index(ascending = True , axis = 1) onetwodcab

5	1.0
6	2.0
7	NaN
8	4.0

sort_index僅有ascending，axis兩個參數(shù)使用方法和按值排序類似

六、OMG太牛辣——DataFrame時間序列

1.pandas時間對象處理

# pd.to_datetime(list)支持批量的字符串轉(zhuǎn)化為時間對象，并且支持各式各樣的時間書寫方式 pd.to_datetime(["2021-01-10","2021/MAY/1"]) DatetimeIndex(['2021-01-10', '2021-05-01'], dtype='datetime64[ns]', freq=None)

2.pandas時間對象自動生成

函數(shù)pd.date_range()的start,end/periods參數(shù)說明

start：開始時間 end：結(jié)束時間 periods：可以指定start不指定end改指定periods，指的是生成從start開始的periods天時間，同樣的，可以指定end不指定start改指定periods # 用start和end生成時間 pd.date_range('2021-01-01','2021-03-01') DatetimeIndex(['2021-01-01', '2021-01-02', '2021-01-03', '2021-01-04','2021-01-05', '2021-01-06', '2021-01-07', '2021-01-08','2021-01-09', '2021-01-10', '2021-01-11', '2021-01-12','2021-01-13', '2021-01-14', '2021-01-15', '2021-01-16','2021-01-17', '2021-01-18', '2021-01-19', '2021-01-20','2021-01-21', '2021-01-22', '2021-01-23', '2021-01-24','2021-01-25', '2021-01-26', '2021-01-27', '2021-01-28','2021-01-29', '2021-01-30', '2021-01-31', '2021-02-01','2021-02-02', '2021-02-03', '2021-02-04', '2021-02-05','2021-02-06', '2021-02-07', '2021-02-08', '2021-02-09','2021-02-10', '2021-02-11', '2021-02-12', '2021-02-13','2021-02-14', '2021-02-15', '2021-02-16', '2021-02-17','2021-02-18', '2021-02-19', '2021-02-20', '2021-02-21','2021-02-22', '2021-02-23', '2021-02-24', '2021-02-25','2021-02-26', '2021-02-27', '2021-02-28', '2021-03-01'],dtype='datetime64[ns]', freq='D') # 用strat和periods生成時間，periods不能省略 pd.date_range('2021-01-01',periods=60) DatetimeIndex(['2021-01-01', '2021-01-02', '2021-01-03', '2021-01-04','2021-01-05', '2021-01-06', '2021-01-07', '2021-01-08','2021-01-09', '2021-01-10', '2021-01-11', '2021-01-12','2021-01-13', '2021-01-14', '2021-01-15', '2021-01-16','2021-01-17', '2021-01-18', '2021-01-19', '2021-01-20','2021-01-21', '2021-01-22', '2021-01-23', '2021-01-24','2021-01-25', '2021-01-26', '2021-01-27', '2021-01-28','2021-01-29', '2021-01-30', '2021-01-31', '2021-02-01','2021-02-02', '2021-02-03', '2021-02-04', '2021-02-05','2021-02-06', '2021-02-07', '2021-02-08', '2021-02-09','2021-02-10', '2021-02-11', '2021-02-12', '2021-02-13','2021-02-14', '2021-02-15', '2021-02-16', '2021-02-17','2021-02-18', '2021-02-19', '2021-02-20', '2021-02-21','2021-02-22', '2021-02-23', '2021-02-24', '2021-02-25','2021-02-26', '2021-02-27', '2021-02-28', '2021-03-01'],dtype='datetime64[ns]', freq='D')

函數(shù)pd.date_range()的freq參數(shù)說明：

freq：指定生成的時間間隔單位，默認為D（天），此外，還有‘H’（小時），‘W’（周）等 W（周）分為，‘W-MON’表示從start開始輸出每周周一，默認只輸入W表示'W-SUN' B：只輸出工作日 # 生成從2021-01-01 00:00:00往后的60小時，每小時生成一個 pd.date_range('2021-01-01',periods=60,freq='H') DatetimeIndex(['2021-01-01 00:00:00', '2021-01-01 01:00:00','2021-01-01 02:00:00', '2021-01-01 03:00:00','2021-01-01 04:00:00', '2021-01-01 05:00:00','2021-01-01 06:00:00', '2021-01-01 07:00:00','2021-01-01 08:00:00', '2021-01-01 09:00:00','2021-01-01 10:00:00', '2021-01-01 11:00:00','2021-01-01 12:00:00', '2021-01-01 13:00:00','2021-01-01 14:00:00', '2021-01-01 15:00:00','2021-01-01 16:00:00', '2021-01-01 17:00:00','2021-01-01 18:00:00', '2021-01-01 19:00:00','2021-01-01 20:00:00', '2021-01-01 21:00:00','2021-01-01 22:00:00', '2021-01-01 23:00:00','2021-01-02 00:00:00', '2021-01-02 01:00:00','2021-01-02 02:00:00', '2021-01-02 03:00:00','2021-01-02 04:00:00', '2021-01-02 05:00:00','2021-01-02 06:00:00', '2021-01-02 07:00:00','2021-01-02 08:00:00', '2021-01-02 09:00:00','2021-01-02 10:00:00', '2021-01-02 11:00:00','2021-01-02 12:00:00', '2021-01-02 13:00:00','2021-01-02 14:00:00', '2021-01-02 15:00:00','2021-01-02 16:00:00', '2021-01-02 17:00:00','2021-01-02 18:00:00', '2021-01-02 19:00:00','2021-01-02 20:00:00', '2021-01-02 21:00:00','2021-01-02 22:00:00', '2021-01-02 23:00:00','2021-01-03 00:00:00', '2021-01-03 01:00:00','2021-01-03 02:00:00', '2021-01-03 03:00:00','2021-01-03 04:00:00', '2021-01-03 05:00:00','2021-01-03 06:00:00', '2021-01-03 07:00:00','2021-01-03 08:00:00', '2021-01-03 09:00:00','2021-01-03 10:00:00', '2021-01-03 11:00:00'],dtype='datetime64[ns]', freq='H') # 生成從2021-01-01 往后的60個周日 pd.date_range('2021-01-01',periods=60,freq='W') DatetimeIndex(['2021-01-03', '2021-01-10', '2021-01-17', '2021-01-24','2021-01-31', '2021-02-07', '2021-02-14', '2021-02-21','2021-02-28', '2021-03-07', '2021-03-14', '2021-03-21','2021-03-28', '2021-04-04', '2021-04-11', '2021-04-18','2021-04-25', '2021-05-02', '2021-05-09', '2021-05-16','2021-05-23', '2021-05-30', '2021-06-06', '2021-06-13','2021-06-20', '2021-06-27', '2021-07-04', '2021-07-11','2021-07-18', '2021-07-25', '2021-08-01', '2021-08-08','2021-08-15', '2021-08-22', '2021-08-29', '2021-09-05','2021-09-12', '2021-09-19', '2021-09-26', '2021-10-03','2021-10-10', '2021-10-17', '2021-10-24', '2021-10-31','2021-11-07', '2021-11-14', '2021-11-21', '2021-11-28','2021-12-05', '2021-12-12', '2021-12-19', '2021-12-26','2022-01-02', '2022-01-09', '2022-01-16', '2022-01-23','2022-01-30', '2022-02-06', '2022-02-13', '2022-02-20'],dtype='datetime64[ns]', freq='W-SUN') # 生成從2021-01-01 往后的60個周五 pd.date_range('2021-01-01',periods=60,freq='W-Fri') DatetimeIndex(['2021-01-01', '2021-01-08', '2021-01-15', '2021-01-22','2021-01-29', '2021-02-05', '2021-02-12', '2021-02-19','2021-02-26', '2021-03-05', '2021-03-12', '2021-03-19','2021-03-26', '2021-04-02', '2021-04-09', '2021-04-16','2021-04-23', '2021-04-30', '2021-05-07', '2021-05-14','2021-05-21', '2021-05-28', '2021-06-04', '2021-06-11','2021-06-18', '2021-06-25', '2021-07-02', '2021-07-09','2021-07-16', '2021-07-23', '2021-07-30', '2021-08-06','2021-08-13', '2021-08-20', '2021-08-27', '2021-09-03','2021-09-10', '2021-09-17', '2021-09-24', '2021-10-01','2021-10-08', '2021-10-15', '2021-10-22', '2021-10-29','2021-11-05', '2021-11-12', '2021-11-19', '2021-11-26','2021-12-03', '2021-12-10', '2021-12-17', '2021-12-24','2021-12-31', '2022-01-07', '2022-01-14', '2022-01-21','2022-01-28', '2022-02-04', '2022-02-11', '2022-02-18'],dtype='datetime64[ns]', freq='W-FRI') # 生成從2021-01-01 往后的60個工作日 pd.date_range('2021-01-01',periods=60,freq='B') DatetimeIndex(['2021-01-01', '2021-01-04', '2021-01-05', '2021-01-06','2021-01-07', '2021-01-08', '2021-01-11', '2021-01-12','2021-01-13', '2021-01-14', '2021-01-15', '2021-01-18','2021-01-19', '2021-01-20', '2021-01-21', '2021-01-22','2021-01-25', '2021-01-26', '2021-01-27', '2021-01-28','2021-01-29', '2021-02-01', '2021-02-02', '2021-02-03','2021-02-04', '2021-02-05', '2021-02-08', '2021-02-09','2021-02-10', '2021-02-11', '2021-02-12', '2021-02-15','2021-02-16', '2021-02-17', '2021-02-18', '2021-02-19','2021-02-22', '2021-02-23', '2021-02-24', '2021-02-25','2021-02-26', '2021-03-01', '2021-03-02', '2021-03-03','2021-03-04', '2021-03-05', '2021-03-08', '2021-03-09','2021-03-10', '2021-03-11', '2021-03-12', '2021-03-15','2021-03-16', '2021-03-17', '2021-03-18', '2021-03-19','2021-03-22', '2021-03-23', '2021-03-24', '2021-03-25'],dtype='datetime64[ns]', freq='B') # 2021-01-01 00:00:00每1小時20分鐘輸出一個時間戳 pd.date_range('2021-01-01',periods=60,freq='1h20min') DatetimeIndex(['2021-01-01 00:00:00', '2021-01-01 01:20:00','2021-01-01 02:40:00', '2021-01-01 04:00:00','2021-01-01 05:20:00', '2021-01-01 06:40:00','2021-01-01 08:00:00', '2021-01-01 09:20:00','2021-01-01 10:40:00', '2021-01-01 12:00:00','2021-01-01 13:20:00', '2021-01-01 14:40:00','2021-01-01 16:00:00', '2021-01-01 17:20:00','2021-01-01 18:40:00', '2021-01-01 20:00:00','2021-01-01 21:20:00', '2021-01-01 22:40:00','2021-01-02 00:00:00', '2021-01-02 01:20:00','2021-01-02 02:40:00', '2021-01-02 04:00:00','2021-01-02 05:20:00', '2021-01-02 06:40:00','2021-01-02 08:00:00', '2021-01-02 09:20:00','2021-01-02 10:40:00', '2021-01-02 12:00:00','2021-01-02 13:20:00', '2021-01-02 14:40:00','2021-01-02 16:00:00', '2021-01-02 17:20:00','2021-01-02 18:40:00', '2021-01-02 20:00:00','2021-01-02 21:20:00', '2021-01-02 22:40:00','2021-01-03 00:00:00', '2021-01-03 01:20:00','2021-01-03 02:40:00', '2021-01-03 04:00:00','2021-01-03 05:20:00', '2021-01-03 06:40:00','2021-01-03 08:00:00', '2021-01-03 09:20:00','2021-01-03 10:40:00', '2021-01-03 12:00:00','2021-01-03 13:20:00', '2021-01-03 14:40:00','2021-01-03 16:00:00', '2021-01-03 17:20:00','2021-01-03 18:40:00', '2021-01-03 20:00:00','2021-01-03 21:20:00', '2021-01-03 22:40:00','2021-01-04 00:00:00', '2021-01-04 01:20:00','2021-01-04 02:40:00', '2021-01-04 04:00:00','2021-01-04 05:20:00', '2021-01-04 06:40:00'],dtype='datetime64[ns]', freq='80T')

3.時間序列

什么是時間序列

時間序列就是以時間對象為索引的Series或DataFrame。eg:

# 創(chuàng)建一個Series，其中index為時間對象 sr = pd.Series(np.arange(1000),index=pd.date_range('2020-01-01',periods=1000)) sr 2020-01-01 0 2020-01-02 1 2020-01-03 2 2020-01-04 3 2020-01-05 4... 2022-09-22 995 2022-09-23 996 2022-09-24 997 2022-09-25 998 2022-09-26 999 Freq: D, Length: 1000, dtype: int32 # 查看sr的index屬性，index確實為時間對象，那么稱sr為一個時間序列 sr.index DatetimeIndex(['2020-01-01', '2020-01-02', '2020-01-03', '2020-01-04','2020-01-05', '2020-01-06', '2020-01-07', '2020-01-08','2020-01-09', '2020-01-10',...'2022-09-17', '2022-09-18', '2022-09-19', '2022-09-20','2022-09-21', '2022-09-22', '2022-09-23', '2022-09-24','2022-09-25', '2022-09-26'],dtype='datetime64[ns]', length=1000, freq='D')

時間序列的特殊作用

時間序列可以直接查找某一年/月/日的數(shù)據(jù) 甚至支持年月日的切片

# 查找2020年所有數(shù)據(jù) sr['2020'] 2020-01-01 0 2020-01-02 1 2020-01-03 2 2020-01-04 3 2020-01-05 4... 2020-12-27 361 2020-12-28 362 2020-12-29 363 2020-12-30 364 2020-12-31 365 Freq: D, Length: 366, dtype: int32 # 查找2020年3月所有數(shù)據(jù) sr['2020-3'] 2020-03-01 60 2020-03-02 61 2020-03-03 62 2020-03-04 63 2020-03-05 64 2020-03-06 65 2020-03-07 66 2020-03-08 67 2020-03-09 68 2020-03-10 69 2020-03-11 70 2020-03-12 71 2020-03-13 72 2020-03-14 73 2020-03-15 74 2020-03-16 75 2020-03-17 76 2020-03-18 77 2020-03-19 78 2020-03-20 79 2020-03-21 80 2020-03-22 81 2020-03-23 82 2020-03-24 83 2020-03-25 84 2020-03-26 85 2020-03-27 86 2020-03-28 87 2020-03-29 88 2020-03-30 89 2020-03-31 90 Freq: D, dtype: int32 # 查找2020年3月19號所有數(shù)據(jù) sr['2020-3-19'] 78 # 查找2020年3月到2021年5月1號所有數(shù)據(jù) sr['2020-03':'2021-05-1'] 2020-03-01 60 2020-03-02 61 2020-03-03 62 2020-03-04 63 2020-03-05 64... 2021-04-27 482 2021-04-28 483 2021-04-29 484 2021-04-30 485 2021-05-01 486 Freq: D, Length: 427, dtype: int32

涉及時間序列的函數(shù)：resample（）——強大的統(tǒng)計函數(shù)

#### resample（）傳入的參數(shù)date_range()的freq參數(shù)相同，比如‘W’可以理解為將所有數(shù)據(jù)按周分組，結(jié)合sum（），mean（）等函數(shù)可以實現(xiàn)數(shù)據(jù)統(tǒng)計 sr 2020-01-01 0 2020-01-02 1 2020-01-03 2 2020-01-04 3 2020-01-05 4... 2022-09-22 995 2022-09-23 996 2022-09-24 997 2022-09-25 998 2022-09-26 999 Freq: D, Length: 1000, dtype: int32 # 統(tǒng)計每周的數(shù)據(jù)總和 sr.resample('W').sum() 2020-01-05 10 2020-01-12 56 2020-01-19 105 2020-01-26 154 2020-02-02 203... 2022-09-04 6818 2022-09-11 6867 2022-09-18 6916 2022-09-25 6965 2022-10-02 999 Freq: W-SUN, Length: 144, dtype: int32 # 統(tǒng)計每周的數(shù)據(jù)平均 sr.resample('W').mean() 2020-01-05 2 2020-01-12 8 2020-01-19 15 2020-01-26 22 2020-02-02 29... 2022-09-04 974 2022-09-11 981 2022-09-18 988 2022-09-25 995 2022-10-02 999 Freq: W-SUN, Length: 144, dtype: int32

七、文件操作

# 讀取csv文件 pd.read_csv('maotai.csv') 日期收盤開盤高低交易量漲跌幅01234...239240241242243

2021/11/12	1,773.78	1,778.00	1,785.05	1,767.00	1.76M	0.24%
2021/11/11	1,769.60	1,752.93	1,769.60	1,741.50	2.27M	0.89%
2021/11/10	1,753.99	1,790.01	1,795.00	1,735.00	3.53M	-2.01%
2021/11/9	1,790.01	1,819.98	1,827.87	1,782.00	2.74M	-1.65%
2021/11/8	1,820.10	1,820.00	1,830.80	1,802.05	1.77M	0.01%
...	...	...	...	...	...	...
2020/11/18	1,693.65	1,715.00	1,720.53	1,683.16	3.52M	-1.29%
2020/11/17	1,715.80	1,740.00	1,742.35	1,701.07	2.52M	-0.82%
2020/11/16	1,730.05	1,711.00	1,730.05	1,697.26	3.06M	1.47%
2020/11/13	1,705.00	1,724.00	1,728.88	1,691.00	2.82M	-1.72%
2020/11/12	1,734.79	1,730.01	1,750.00	1,722.27	2.35M	0.20%

244 rows × 7 columns

# 參數(shù)index_col：指定索引，可以傳數(shù)字表示第n列，也可以傳列名 df = pd.read_csv('maotai.csv',index_col=0) df 收盤開盤高低交易量漲跌幅日期2021/11/122021/11/112021/11/102021/11/92021/11/8...2020/11/182020/11/172020/11/162020/11/132020/11/12

1,773.78	1,778.00	1,785.05	1,767.00	1.76M	0.24%
1,769.60	1,752.93	1,769.60	1,741.50	2.27M	0.89%
1,753.99	1,790.01	1,795.00	1,735.00	3.53M	-2.01%
1,790.01	1,819.98	1,827.87	1,782.00	2.74M	-1.65%
1,820.10	1,820.00	1,830.80	1,802.05	1.77M	0.01%
...	...	...	...	...	...
1,693.65	1,715.00	1,720.53	1,683.16	3.52M	-1.29%
1,715.80	1,740.00	1,742.35	1,701.07	2.52M	-0.82%
1,730.05	1,711.00	1,730.05	1,697.26	3.06M	1.47%
1,705.00	1,724.00	1,728.88	1,691.00	2.82M	-1.72%
1,734.79	1,730.01	1,750.00	1,722.27	2.35M	0.20%

244 rows × 6 columns

# 默認情況下生成的列名均為字符串格式 df.index Index(['2021/11/12', '2021/11/11', '2021/11/10', '2021/11/9', '2021/11/8','2021/11/5', '2021/11/4', '2021/11/3', '2021/11/2', '2021/11/1',...'2020/11/25', '2020/11/24', '2020/11/23', '2020/11/20', '2020/11/19','2020/11/18', '2020/11/17', '2020/11/16', '2020/11/13', '2020/11/12'],dtype='object', name='日期', length=244) # 參數(shù)parse_dates=True將所有能用時間對象表示的列統(tǒng)統(tǒng)轉(zhuǎn)為時間對象 df = pd.read_csv('maotai.csv',index_col=0,thousands=',',parse_dates=True) df 收盤開盤高低交易量漲跌幅日期2021-11-122021-11-112021-11-102021-11-092021-11-08...2020-11-182020-11-172020-11-162020-11-132020-11-12

1773.78	1778.00	1785.05	1767.00	1.76M	0.24%
1769.60	1752.93	1769.60	1741.50	2.27M	0.89%
1753.99	1790.01	1795.00	1735.00	3.53M	-2.01%
1790.01	1819.98	1827.87	1782.00	2.74M	-1.65%
1820.10	1820.00	1830.80	1802.05	1.77M	0.01%
...	...	...	...	...	...
1693.65	1715.00	1720.53	1683.16	3.52M	-1.29%
1715.80	1740.00	1742.35	1701.07	2.52M	-0.82%
1730.05	1711.00	1730.05	1697.26	3.06M	1.47%
1705.00	1724.00	1728.88	1691.00	2.82M	-1.72%
1734.79	1730.01	1750.00	1722.27	2.35M	0.20%

244 rows × 6 columns

df.index DatetimeIndex(['2021-11-12', '2021-11-11', '2021-11-10', '2021-11-09','2021-11-08', '2021-11-05', '2021-11-04', '2021-11-03','2021-11-02', '2021-11-01',...'2020-11-25', '2020-11-24', '2020-11-23', '2020-11-20','2020-11-19', '2020-11-18', '2020-11-17', '2020-11-16','2020-11-13', '2020-11-12'],dtype='datetime64[ns]', name='日期', length=244, freq=None) # 參數(shù)parse_dates=[n]將第n列列統(tǒng)統(tǒng)轉(zhuǎn)為時間對象 df = pd.read_csv('maotai.csv',index_col=0,parse_dates=[0]) df.index DatetimeIndex(['2021-11-12', '2021-11-11', '2021-11-10', '2021-11-09','2021-11-08', '2021-11-05', '2021-11-04', '2021-11-03','2021-11-02', '2021-11-01',...'2020-11-25', '2020-11-24', '2020-11-23', '2020-11-20','2020-11-19', '2020-11-18', '2020-11-17', '2020-11-16','2020-11-13', '2020-11-12'],dtype='datetime64[ns]', name='日期', length=244, freq=None) # header=None:不讓csv文件中的第0行成為列名 pd.read_csv('maotai.csv',index_col=0,header=None) 1234560日期2021/11/122021/11/112021/11/102021/11/9...2020/11/182020/11/172020/11/162020/11/132020/11/12

收盤	開盤	高	低	交易量	漲跌幅
1,773.78	1,778.00	1,785.05	1,767.00	1.76M	0.24%
1,769.60	1,752.93	1,769.60	1,741.50	2.27M	0.89%
1,753.99	1,790.01	1,795.00	1,735.00	3.53M	-2.01%
1,790.01	1,819.98	1,827.87	1,782.00	2.74M	-1.65%
...	...	...	...	...	...
1,693.65	1,715.00	1,720.53	1,683.16	3.52M	-1.29%
1,715.80	1,740.00	1,742.35	1,701.07	2.52M	-0.82%
1,730.05	1,711.00	1,730.05	1,697.26	3.06M	1.47%
1,705.00	1,724.00	1,728.88	1,691.00	2.82M	-1.72%
1,734.79	1,730.01	1,750.00	1,722.27	2.35M	0.20%

245 rows × 6 columns

# names:自行指定列名 pd.read_csv('maotai.csv',index_col=0,header=None,names=['a','b','c','d','e','f','g']) bcdefga日期2021/11/122021/11/112021/11/102021/11/9...2020/11/182020/11/172020/11/162020/11/132020/11/12

收盤	開盤	高	低	交易量	漲跌幅
1,773.78	1,778.00	1,785.05	1,767.00	1.76M	0.24%
1,769.60	1,752.93	1,769.60	1,741.50	2.27M	0.89%
1,753.99	1,790.01	1,795.00	1,735.00	3.53M	-2.01%
1,790.01	1,819.98	1,827.87	1,782.00	2.74M	-1.65%
...	...	...	...	...	...
1,693.65	1,715.00	1,720.53	1,683.16	3.52M	-1.29%
1,715.80	1,740.00	1,742.35	1,701.07	2.52M	-0.82%
1,730.05	1,711.00	1,730.05	1,697.26	3.06M	1.47%
1,705.00	1,724.00	1,728.88	1,691.00	2.82M	-1.72%
1,734.79	1,730.01	1,750.00	1,722.27	2.35M	0.20%

245 rows × 6 columns

# na_values:指定哪些字符串被指定為NaN pd.read_csv('maotai.csv',index_col=0,header=None,names=['a','b','c','d','e','f','g'],na_values=['收盤','開盤']) bcdefga日期2021/11/122021/11/112021/11/102021/11/9...2020/11/182020/11/172020/11/162020/11/132020/11/12

NaN	NaN	高	低	交易量	漲跌幅
1,773.78	1,778.00	1,785.05	1,767.00	1.76M	0.24%
1,769.60	1,752.93	1,769.60	1,741.50	2.27M	0.89%
1,753.99	1,790.01	1,795.00	1,735.00	3.53M	-2.01%
1,790.01	1,819.98	1,827.87	1,782.00	2.74M	-1.65%
...	...	...	...	...	...
1,693.65	1,715.00	1,720.53	1,683.16	3.52M	-1.29%
1,715.80	1,740.00	1,742.35	1,701.07	2.52M	-0.82%
1,730.05	1,711.00	1,730.05	1,697.26	3.06M	1.47%
1,705.00	1,724.00	1,728.88	1,691.00	2.82M	-1.72%
1,734.79	1,730.01	1,750.00	1,722.27	2.35M	0.20%

245 rows × 6 columns

# df.to_文件格式（）保存為某個文件格式

總結(jié)

以上是生活随笔為你收集整理的成为华尔街金融巨鳄第三课： Pandas2:学会使用Pandas-DataFrame的全部內(nèi)容，希望文章能夠幫你解決所遇到的問題。

如果覺得生活随笔網(wǎng)站內(nèi)容還不錯，歡迎將生活随笔推薦給好友。