Could be a related post below. Looks like MarketStore supports web socket.

https://community.backtrader.com/topic/1120/marketstore-new-data-feed-complete-code-example

Websocket would make backtrader more awesome! And it would be of great help if someone could help provide some working examples.

@bb2 The issue seems to be gone in latest version.

https://community.backtrader.com/topic/1397/resampling-fixes-for-1-9-67-122

@pavelan Just a wild guess, please don't take my words without testing. Backtrader is using "rightedge=True" by default, meaning at timestamp "2018-09-18T00:00:00" you are seeing the information for the period "2018-09-17T00:00:01" to "2018-09-17T23:59:59" roughly. In other words, "2018-09-18T00:00:00" marks the end of "2018-09-17".

While on most other places like Bloomberg or even InteractiveBroker, a timestamp "2018-09-18" marks the beginning of a bar, meaning that the OHLC for the bar associated with timestamp "2018-09-18T00:00:00" is for what has happened during the day time of "2018-09-18".

@gwadaman Did you managed to find out more? I am looking for the same too.

@backtrader This is awesome! Thanks for the quick fix.

@ab_trader thanks for the tip, I will try it later. My plan is to trade mupltie securities simultaneously but I've experienced a lot of unexpected issues during live trading with only one security (ibstore issues, resampling issues, connection issues, IBC issues etc).

Living trading is a totally different beast from back testing. So I just want to cheat a bit and see whether anyone could share some experience.

The code is not quite stable as sometimes it would return an error like below, sometimes it would run just fine.

***** DATA NOTIF: DELAYED
***** DATA NOTIF: DELAYED
***** DATA NOTIF: DELAYED
***** DATA NOTIF: DELAYED
***** DATA NOTIF: UNKNOWN -420
Traceback (most recent call last):
.......
return self.data.datetime.datetime(0)
File ".../backtrader/linebuffer.py", line 387, in datetime
return num2date(self.array[self.idx + ago],
IndexError: array index out of range

Hello
I have two 1 minute bar live data feeds from ibstore, one for SPX 500 E-mini and the other for USDJPY currency. The former is subject to some fixed trading halt and maintenance hours on a daily basis, the latter has no such halt hours. Will Cerebro continue to run as usual or will it stop when it sees one data stops coming? If it stops, is there any work around to make it run continuously?

Just want to know how Cerebro keeps multiple datas in sync in live trading. I couldn't find much previous related discussion so I'm posting the question here.

Many thanks,

Kevin

Hello

Suppose I want live data for a same security in 4 different timeframes: 1 minute, 15 minutes, 1 hour and 1 day, and trade on the smallest granular data which is 1 minute bar. I also need to backfill from a specific historical start date to make sure good indicator values coming out from the daily bars. To make sure I don't miss any good minute bar data, the smallest granular data source is 20 seconds bar with "qcheck=2" and "rtbar=True". Is below idiomatic way to do it? Do I have to specify "fromdate" argument for every "ibstore.getdata(...)" line?

Many thanks.

Kevin

data0= ibstore.getdata('FOO', timeframe=bt.TimeFrame.Seconds, compression=20, 
                    rtbar=True,
                    qcheck =2,
                    fromdate=datetime.datetime(2018,9,1) )

min_bar= cerebro.resampledata(data0, timeframe=bt.TimeFrame.Minutes,compression=1,name='min_bar')

data1= ibstore.getdata('FOO', timeframe=bt.TimeFrame.Minutes,  compression=15,
                    fromdate=datetime.datetime(2018,9,1))    

quarter_bar= cerebro.resampledata(data1, timeframe=bt.TimeFrame.Minutes,
                    compression=15,name='quarter_bar')    

data2= ibstore.getdata('FOO', timeframe=bt.TimeFrame.Minutes,  compression=60, 
                    fromdate=datetime.datetime(2018,9,1))    

hr_bar= cerebro.resampledata(data2, timeframe=bt.TimeFrame.Minutes,
                    compression=60,name='hr_bar')    

data3=ibstore.getdata(data2,timeframe=bt.TimeFrame.Days,compression=1,
                    fromdate=datetime.datetime(2018,9,1))   

day_bar= cerebro.resampledata(data3, timeframe=bt.TimeFrame.Days,
                    compression=1,name='day_bar')