Is there a listing of when OnData is called?

Hi Mak;

Here is a good example. In this example nothing happens for 30 days because the data is not ready on the first OnData call and OnData is not called again for 30 days. This is from boot camp “Algorithmic Trading Using Python #6 - Trading & Orders”

from collections import deque

class AdaptableSkyBlueHornet(QCAlgorithm):

    def Initialize(self):
        self.SetStartDate(2020, 1, 1)
        self.SetEndDate(2021, 1, 1)
        self.SetCash(100000)
        self.spy = self.AddEquity("SPY", Resolution.Daily).Symbol
        
    #    self.sma = self.SMA(self.spy, 30, Resolution.Daily)
        
        # History warm up for shortcut helper SMA indicator
    #    closing_prices = self.History(self.spy, 30, Resolution.Daily)["close"]
    #    for time, price in closing_prices.loc[self.spy].items():
    #        self.sma.Update(time, price)
        
        # Custom SMA indicator
        self.sma = CustomSimpleMovingAverage("CustomSMA", 30)
        self.RegisterIndicator(self.spy, self.sma, Resolution.Daily)

    
    def OnData(self, data):
        if not self.sma.IsReady:
            return
        
        # Save high, low, and current price
        hist = self.History(self.spy, timedelta(365), Resolution.Daily)
        low = min(hist["low"])
        high = max(hist["high"])
        
        price = self.Securities[self.spy].Price
        
        # Go long if near high and uptrending
        if price * 1.05 >= high and self.sma.Current.Value < price:
            if not self.Portfolio[self.spy].IsLong:
                self.SetHoldings(self.spy, 1)
        
        # Go short if near low and downtrending
        elif price * 0.95 <= low and self.sma.Current.Value > price:  
            if not self.Portfolio[self.spy].IsShort:
                self.SetHoldings(self.spy, -1)
        
        # Otherwise, go flat
        else:
            self.Liquidate()
        
        self.Plot("Benchmark", "52w-High", high)
        self.Plot("Benchmark", "52w-Low", low)
        self.Plot("Benchmark", "SMA", self.sma.Current.Value)


class CustomSimpleMovingAverage(PythonIndicator):
    
    def __init__(self, name, period):
        self.Name = name
        self.Time = datetime.min
        self.Value = 0
        self.queue = deque(maxlen=period)

    def Update(self, input):
        self.queue.appendleft(input.Close)
        self.Time = input.EndTime
        count = len(self.queue)
        self.Value = sum(self.queue) / count
        # returns true if ready
        return (count == self.queue.maxlen)

Hi Mak;

Here is another example - from “Algorithmic Trading Using Python #7 - Consolidators and Rolling Windows”. In this code it has:

class GapReversalAlgo(QCAlgorithm):

    def Initialize(self):
        self.SetStartDate(2018, 1, 1)
        self.SetEndDate(2021, 1, 1)
        self.SetCash(100000)
        self.symbol = self.AddEquity("SPY", Resolution.Minute).Symbol
        self.rollingWindow = RollingWindow[TradeBar](2)
        self.Consolidate(self.symbol, Resolution.Daily, self.CustomBarHandler)
        
        self.Schedule.On(self.DateRules.EveryDay(self.symbol),
                 self.TimeRules.BeforeMarketClose(self.symbol, 15),      
                 self.ExitPositions)

    def OnData(self, data):
        
        if not self.rollingWindow.IsReady:
            return
        
        if not (self.Time.hour == 9 and self.Time.minute == 31):
            return
        
        # Gap Up => Sell
        if data[self.symbol].Open >= 1.01*self.rollingWindow[0].Close:
            self.SetHoldings(self.symbol, -1)
        # Gap Down => Buy
        elif data[self.symbol].Open <= 0.99*self.rollingWindow[0].Close:
            self.SetHoldings(self.symbol, 1)

    def CustomBarHandler(self, bar):
        self.rollingWindow.Add(bar)

    def ExitPositions(self):
        self.Liquidate(self.symbol)

The above code depends on OnData() being called at exactly 9:31. Is this going to happen because of the "self.AddEquity("SPY", Resolution.Minute)"? And is that the purpose of the resolution in AddEquity?

And… what if due to delays in other methods called, by the time it calls this, it's now 9:32. Can that happen? Or is self.Time set to what the time should be when this is called?

thanks - dave

OnData is run every time your algorithm receives new data, so this depends on what resolution your data is subscribed to. Resolution.Minute will make OnData run after every minute closes, Resolution.Hour after every time the hour closes, etc. Alternatively, you could also set up a scheduled event in Initialize() so you can get code to run exactly when you want everyday.

Also, simply warm up your algorithm in Initialize() using self.SetWarmUp(period) so that you don't have to wait 30 days for OnData to run the first time.

Hi Vncne;

Thank you. This is helping me make sense of all this. Finally, I think…

Question 1: Ok, so in a lot of the examples you're trading one stock. And so when the stock updates, that's when it's time to determine if I want to do something. That works great.

But say I have 40 stocks I'm working with, and I'm going to buy the 10 that best fit my criteria. So I need all 10 to update to their latest data (let's say daily and it's after the market closed). What then? 

1. I could set a timer event for ½ hour after market close assuming all updated by then (and verify of course and wait another ½ hour if they haven't).
2. When OnData is called, I mark that stock updated, and then walk all 40. If all 40 have been updated, then I do my nightly processing.
3. Another approach???

Question 2: Working my way through the boot camp videos, they tend to mention self.SetWarmUp(period) but then not use it. So I was figuring that best practices was not call anything in the methods that could block. Is that wrong? Should I always use self.SetWarmUp(period) so I know my data is all read in?

thanks - dave

Hi David Thielen,
The Ondata method is the primary event handler for securities in the Universe; it is called whenever there is data for the subscribed security(trade and non-trade data), so in case of daily resolution, it would be called at midnight when the day has ended and you could iterate through the data to scan for signals and any order placed would be converted to OnMarketOpen order type. We do not need to explicitly check if our indicators are ready we can have an Updated method to receive updates whenever to scan for insights.You can either use self.SetWarmUp or use history request to make your indicators ready. It is always a good practice to check whether an indicator is ready before using it:
 

if not self.ema.IsReady:
   return

Best,
Varad Kabade