I was just saying I did the loop different using 'for' instead of enumerate
but I do something like this

params = ('portfolio', ['AAPL', 'GOOG', 'SPY', 'IWM', 'EEM'])    

def next(self):
    for symbol in self.p.portfolio:

        d = self.getdatabyname(symbol)
        if self.order[symbol]:
            continue
       
       if self.signal:
            self.order[symbol] = self.buy(d, size=self.trade_size[symbol])

@cwse
I create the self.order dict in init and don't have a problem
you would have to run it in debug and figure out what key it fails on

the only thing I could see is that my loop is different as I pass all my symbols as param called portfolio and then just do loop that is

for symbol in self.p.portfolio:

@cwse
I use the following for multiple symbols

in your init where you define self.order = None, make it a dict , self.order = {}

def __init__(self):
    self.order = {}
    for i, d in enumerate(d for d in self.datas if len(d)):
       self.order[d._name] = None

def notify(self, order):
# standard code for notification 
    self.order[order.data._name] = None

def next(self):
   for i, d in enumerate(d for d in self.datas if len(d)):
       if not self.order[d._name]:
          continue

try this:
cerebro.addwriter(bt.WriterFile, csv = True, out='your_strategy_results')

if you are using IB, you should be able to put each system in a sub account and manage the positions this way.

you can also access the IB transaction log and I think you can retrieve things by order id, but I know this was somewhat of an issue for us, IB was not all that helpful. on a daily basis you should probably reconcile what your system thinks your position is vs what the broker has, the brokers make mistakes more often than you would think

we had a similar issue with needing persistent state variables for live trading and ended up having to store them in a database, but I would think you could do the same in a single csv file and just read the last lines in at startup.

note we solved it by creating a new cerebro class that lets us do the following

cerebro.add_feeds(startdate, enddate, portfolio_symbols, conflations)

not sure how to get the formatted code to paste

startdate = datetime.datetime(2012, 1, 1)
enddate = datetime.datetime(2013, 1, 3)
def get_mstar_data(sym, startdate, enddate):
    s = startdate.year
    e = enddate.year
    df = None
    for i in range(s, e + 1):
        for filename in glob.glob("%s/%d/%s*_1min.gz" % (m_star_cbot_path, i, sym)):
            filedf = _create_df(filename)
            filedf = filedf.ix[startdate:enddate]
            print (len(filedf.index))
            df = filedf if df is None else df.append(filedf)



    return mstar(
        fromdate=startdate,
        todate=enddate,
        dataname=df,
        name=sym,
        datetime=None

    )


data = get_mstar_data('ZC0Y', startdate, enddate):

cerebro.adddata(data)
cerebro.resampledata(data, name=ZC0Y15m, timeframe=bt.Timeframe.Minutes, compression=15)

not sure if anyone else has had trouble with loading data from a pandas dataframe and then using cerebro.resample, but we did. The dataframe for minute data loads find and then when I go to pass the dataframte to cerebro.resample I get nan values. if I use a csv it loads and resamples with no problems

so we ended up using pandas own resample feature which works just fine

your daily resampled data will have the previous days price because you don't know todays values
you might want to try and replay the data if you are wanting the daily values to update through out the day

@Ed-Bartosh
the bulk of the time you have 16:00 as you close, 16:15 for some of the index ETFs,
you can download the exchange calendars and then put the holidays in where you early close

if you have some timezone issues then try using pytz to manage that, I think there is some functionality in backtrader for it

another option would be to count minutes from the first trade of the day, but you still have to manage the holidays