所以我有一个csv,里面有微软的分钟股票数据。我试图找到每个交易日的低点。代码如下所示:
ticker='MSFT'
df = pd.read_csv('/Volumes/Seagate Portable/S&P 500 List/{}.txt'.format(ticker))
df.columns = ['Extra', 'Dates', 'Open', 'High', 'Low', 'Close', 'Volume']
df.Dates = pd.to_datetime(df.Dates)
df.set_index(df.Dates, inplace=True)
df.drop(['Extra', 'High', 'Volume', 'Dates', 'Open'], axis=1, inplace=True)
df = df.between_time('9:30', '16:00')
df['Low'] = df.Low.groupby(by=[df.index.day]).min()
df
输出为:
Low Close
Dates
2020-01-02 09:30:00 NaN 158.610
2020-01-02 09:31:00 NaN 158.380
2020-01-02 09:32:00 NaN 158.620
2020-01-02 09:33:00 NaN 158.692
2020-01-02 09:34:00 NaN 158.910
... ... ...
2020-12-18 15:56:00 NaN 218.700
2020-12-18 15:57:00 NaN 218.540
2020-12-18 15:58:00 NaN 218.710
2020-12-18 15:59:00 NaN 218.150
2020-12-18 16:00:00 NaN 218.500
所以问题是,低点充满了NaN值,我之所以要测量是因为我错过了使用groupby。我也尝试过:
ticker='MSFT'
df = pd.read_csv('/Volumes/Seagate Portable/S&P 500 List/{}.txt'.format(ticker))
df.columns = ['Extra', 'Dates', 'Open', 'High', 'Low', 'Close', 'Volume']
df.Dates = pd.to_datetime(df.Dates)
df.set_index(df.Dates, inplace=True)
df.drop(['Extra', 'High', 'Volume', 'Dates', 'Open'], axis=1, inplace=True)
df = df.between_time('9:30', '16:00')
df = df.groupby(by=[df.index.day]).min()
df
其输出为:
Low Close
Dates
1 150.8200 150.9800
2 150.3600 150.8400
3 152.1900 152.2800
4 165.6200 165.7000
5 165.6900 165.8200
6 156.0000 156.0700
7 157.3200 157.3500
8 157.9491 158.0000
9 150.0000 150.2700
10 152.5800 152.7950
11 151.1500 151.1930
12 138.5800 138.7600
13 140.7300 140.8700
14 161.7200 161.7500
15 162.5700 162.6300
16 135.0000 135.3300
17 135.0000 135.3400
18 135.0200 135.2600
19 139.0000 139.1300
20 135.8600 136.5900
21 166.1102 166.2100
22 165.6800 165.6900
23 132.5200 132.7100
24 141.2700 141.6481
25 144.4400 144.8102
26 148.3700 149.7000
27 149.2000 149.2700
28 152.0000 153.8152
29 165.6900 165.7952
30 150.0100 152.7200
31 156.5600 157.0450
问题是,它正在寻找关闭和打开的低点。此外,总共只有31行,不过应该还有更多的行,因为这是一个2020年的数据集。我认为这样做是不对的,因为我看了前31天每天的收盘价,不可能这些都是这些天的低点。所以问题是,我如何才能找到每天的低点,而不影响收尾专栏,并避免上述问题
试试这个:
最后,创建一个新的数据帧:
相关问题 更多 >
编程相关推荐