Questions about Rolling Windows

Hi Nico

A fews things to mention in here:

1. You’re overwriting the indicators and RollingWindows for every date of the SEC historical data History. I’m not sure why you’ll need to initiate a new RollingWindow and STD indicator for each day, at least not from the logic in the attached backtest I saw, so you might want to remove “for date in a” and only do so for each symbol object.
2. Same as above point, you’re overwriting that in every symbol added into the universe. I recommend you to implement a SymbolData class for each symbol added and add the RollingWindows/indicators there. Also, you’ll need a dictionary to hold the <Symbol, SymbolData> pairs. Remember to dispose the object after the symbol is removed.
3. STD indicator is called with the SEC symbol, I guess you‘re actually referring to the ticker symbol? Actually, you don’t need a STD indicator, you can just do a rolling std using pandas in-built function within the history data frame:
   series_std = history_std.close.rolling(10).std()
   Then iterate the last 30 rows in the series.

Best
Louis

Louis Szeto 
You are right, the code is very confusing. I wrote a short code in the research notebook to show how it should look like.
I'm trying to archieve this with rolling windows so i can access the values in a simple manner. The problem is no matter what I'm trying in QC, it wont work. 

for date in report_history.index:
    dates = date.to_pydatetime()
    print(type(dates))
    print("SEC Filling date: " + str(date))
    std = StandardDeviation(20)
    display(self.Indicator(std, msft, dates - timedelta(days=60), dates, Resolution.Daily))

Hey again,

I was able to get the 30(in code 50) day history for each of the days and adding the rolling std to it. Thank you for that.

for date in a:
	self.Debug("SEC Filling Date: " + str(date))
    dates = date.to_pydatetime()
    history2 = self.History(symbol, dates - timedelta(days=50), dates, Resolution.Daily)
    self.window = RollingWindow[TradeBar](30)
    series_std = history2['close'].rolling(20).std()
    self.Debug(series_std)

But how do I add the series_std into the window? I assume I cant just ‘self.window.Add(series_std)’
Then how can I add the data into it?

Hey Louis Szeto 

I was able to figure out another way to get the information I needed. 

day = series_std[symbol].idxmin() - date

But I'm still asking myself if adding the series_std into a rolling window would work. Because it seems that just data that is newly generated and updated can be added into the rolling windows and not a ‘Dataframe’.

Anyways thank you for your help!