How to fill missing data

MikeTheTIke

I tried many ways to get SessionFiller to work...without success:

Data looks like this (with some data missing):

2017 11 19 02:40:00,0.026441389694809914,0.026441389694809914,0.026441389694809914,0.026441389694809914,158.64834594726562
2017 11 19 02:45:00,,,,,
2017 11 19 02:50:00,,,,,
2017 11 19 02:55:00,,,,,
2017 11 19 03:00:00,0.026379242539405823,0.026379242539405823,0.026301197707653046,0.026301197707653046,359.79620361328125

Some sample code:

dtstart = datetime.datetime.strptime('2017 11 18 12:05:00', '%Y %m %d %H:%M:%S')
dtend = datetime.datetime.strptime('2017 12 18 12:05:00', '%Y %m %d %H:%M:%S')

# Create a Data Feed
data = bt.feeds.GenericCSVData(
    timeframe=bt.TimeFrame.Minutes,
    compression=5,
    dataname=datapath,
    openinterest=-1,
    fromdate=dtstart,
    todate=dtend,
    dtformat="%Y %m %d %H:%M:%S",
    sessionstart=dtstart,
    sessionend=dtend,
    )
data.addfilter(bt.filters.SessionFiller, fill_vol=0.0)
cerebro.adddata(data)

I tried many variations, but in none of the plots or logic it seems SessionFiller is actually inserting something.

What's wrong here?

Paska Houso

Your data is not missing ... it's simply empty, but you still have a timestamp ... that may be the key.

MikeTheTIke

That's one of the variations I tried :) I removed the "empty" lines and ran the code above. In the plots the missing timespans are missing as well. I would have expected them to be filled with the last values.

Or is there a special way to access the filled data?