2

I have 1-min ticker information of Apple in a dataframe as so:

             Local time   Open    High     Low   Close  Volume
0   2018-04-19 15:00:00 46.707  46.708  46.687  46.687  0.0150
1   2018-04-19 15:01:00 46.688  46.688  46.667  46.688  0.0320
2   2018-04-19 15:02:00 46.687  46.728  46.677  46.728  0.0091
3   2018-04-19 15:03:00 46.727  46.728  46.708  46.717  0.0332
4   2018-04-19 15:04:00 46.708  46.718  46.677  46.677  0.0243

I have converted the "Local time" column into datetime using pd.to_datetime(df['Local time']). I want to go through each day individually to backtest a strategy. But I do not know how to loop through chunks of the df one at a time defined by a change in date. I tried using some for loops but they did not work since the number of minutes traded is apparently different on some days (not 390):

index = 390 #Number of traded minutes on most days
rows = 286155 #number of rows in the dataset
for x in range(286155/390):
    index = index * x
    index2 = index * (x-1)
    for y in df[index2:index]:
        '''Strategy to be Executed for that day'''

How can I achieve what I want to do?

Aryagm
  • 530
  • 1
  • 7
  • 18

1 Answers1

4

As suggested by @Ben.T:

for dt, df in data.groupby(data["Local time"].dt.date):
    print(f"\n[{dt}]")
    print(df.head())
    # do stuff here

[2021-04-16]
           Local time  Value
0 2021-04-16 00:00:00  28.15
1 2021-04-16 00:01:00  25.33
2 2021-04-16 00:02:00  82.04
3 2021-04-16 00:03:00  17.81
4 2021-04-16 00:04:00  80.71

[2021-04-17]
              Local time  Value
1440 2021-04-17 00:00:00  67.72
1441 2021-04-17 00:01:00  52.91
1442 2021-04-17 00:02:00  26.40
1443 2021-04-17 00:03:00  69.11
1444 2021-04-17 00:04:00  91.88

[2021-04-18]
              Local time  Value
2880 2021-04-18 00:00:00  13.03
2881 2021-04-18 00:01:00  53.42
2882 2021-04-18 00:02:00   9.28
2883 2021-04-18 00:03:00  77.74
2884 2021-04-18 00:04:00  24.91
Corralien
  • 109,409
  • 8
  • 28
  • 52