Forward dates returning values?

run-out

The following code shouldn't be returning forward values but it is.

def __init__(self):
        self.open = self.datas[0].open
        self.close = self.datas[0].close

    def next(self):
        date = self.data.datetime.date()

        if date > datetime.date(2006, 12, 1):
            return

        print(
            date,
            self.datas[0].close[0],
            self.close[3],
            self.datas[0].open[0],
            self.open[3],
        )

Output:

2006-01-02 3604.33 3650.24 3578.73 3652.19
2006-01-03 3614.34 3666.99 3604.08 3650.54
2006-01-04 3652.46 3671.78 3615.23 3667.1
2006-01-05 3650.24 3644.94 3652.19 3671.23
2006-01-06 3666.99 3668.61 3650.54 3645.73
2006-01-09 3671.78 3670.2 3667.1 3667.16
2006-01-10 3644.94 3629.25 3671.23 3670.27
2006-01-11 3668.61 3644.41 3645.73 3628.73
2006-01-12 3670.2 3610.07 3667.16 3639.57

@vladisld Do I need another coffee? What's going on here?

Full Code:

import datetime
import backtrader as bt
import pandas as pd
from extension.data import IQMinute
from dateutil import relativedelta
from datetime import date


class Strategy(bt.Strategy):
    def log(self, txt, dt=None):
        """ Logging function fot this strategy"""
        dt = dt or self.data.datetime[0]
        if isinstance(dt, float):
            dt = bt.num2date(dt)
        print("%s, %s" % (dt.date(), txt))


    def __init__(self):
        self.open = self.datas[0].open
        self.close = self.datas[0].close

    def next(self):
        date = self.data.datetime.date()

        if date > datetime.date(2006, 12, 1):
            return

        print(
            date,
            self.datas[0].close[0],
            self.close[3],
            self.datas[0].open[0],
            self.open[3],
        )

if __name__ == "__main__":
    cerebro = bt.Cerebro()
    data = bt.feeds.GenericCSVData(
        dataname="data/2006-day-001.txt",
        dtformat=("%Y-%m-%d"),
        timeframe=bt.TimeFrame.Days,
        compression=1,
    )
    cerebro.adddata(data)
    cerebro.addstrategy(Strategy)

    cerebro.run()

vladisld

@run-out said in Forward dates returning values?:

if date > datetime.date(2006, 12, 1):
return

I don't see the problem other than probably mixup of months and days. The code prints on each next until Dec'1 2006 - and the output confirms that.

The print seems to be strange - but just probably because it prints in US locale ( %Y-%d-%m ) and not (%Y-%m-%d)

run-out

@vladisld Hey Vlad,

I'm using the conditional this just to stop the error at the end of the dataset, and this is not the important part..

 if date > datetime.date(2006, 12, 1):
            return

The important part is that I getting the values three days ahead of time when I should get an error from backtrader.

 self.datas[0].close[0],
 self.close[3],

Note that the close column is reading ahead 3 days. 3650.24 in the third column is the close value 3 bars ahead in the second column. We are looking into the future. Should backtrader kick up an error?

 2006-01-02 3604.33 3650.24 3578.73 3652.19
2006-01-03 3614.34 3666.99 3604.08 3650.54
2006-01-04 3652.46 3671.78 3615.23 3667.1
2006-01-05 3650.24 3644.94 3652.19 3671.23

vladisld

@run-out said in Forward dates returning values?:

We are looking into the future. Should backtrader kick up an error?

Why should backtrader kick up an error ? Referencing the future values are technically possible (it's questionable logically though)

run-out

@vladisld I guess i haven't done that in so long I came to believe it wasn't possible.