Getting weekly data in next(self) when data source is sparse hourly data

paindemie

My data source is hourly with many hours missing but I only need weekly data in my custom indicator/strategy. I used timeframe=bt.TimeFrame.Weeks when calling bt.feeds.GenericCSVData(..) but I'm still getting every tick from my hourly data in the next() method, where I only want one value a week (same for my strategy where every data point should be weekly), how can I achieve that? Thanks.

backtrader

Docs - Resample

paindemie

I did use timeframe=bt.TimeFrame.Weeks, still getting every point of data, one after another in the next

ab_trader

@paindemie please show the code. if you call correctly cerebro.resampledata(data, **kwargs) it should be weekly bars returned.

paindemie

data = bt.feeds.GenericCSVData(
    dataname=data_source,
    fromdate=start,
    todate=stop,
    nullvalue=0.0,
    dtformat=('%Y-%m-%d %H:%M:%S'),
    tmformat=('%Y-%m-%d %H:%M:%S'),
    openinterest=-1,
    timeframe=bt.TimeFrame.Days,
    compression=1,
    plot=False)


# Create a cerebro entity
cerebro = bt.Cerebro(stdstats=True)

#cerebro.adddata(data)
cerebro.resampledata(data, bt.TimeFrame.Weeks)

in next self.data[-1] is the DAY before self.data[0] where it should be 7 days before if Im not missing smith obvious..

ab_trader

All examples that I've seen use the following syntax:

cerebro.resampledata(data, timeframe=bt.TimeFrame.Weeks)

Maybe resampledata method doesn't recognize timeframe in your case.

backtrader

@paindemie said in Getting weekly data in next(self) when data source is sparse hourly data:

cerebro.resampledata(data, bt.TimeFrame.Weeks)

That won't resample to anything unless you use the right syntax (as pointed out by @ab_trader)

backtrader

@ab_trader said in Getting weekly data in next(self) when data source is sparse hourly data:

Maybe resampledata method doesn't recognize timeframe in your case.

It doesn't because the timeframe named argument is not being provided.

paindemie

Aaah got it thanks