i handled the exception, i`m not sure if this would be valid for real trading though

for stock in context.fundamental_df:
if (stock not in context.portfolio.positions):
#log.info("Ordering %s" % (stock.symbol))
#add/update the stock to the list
try: #workaround for the ordering bug
stock_price = data[stock].price
context.mypositions[stock] = context.days
order(stock,rebalance_budget/stock_price) #use order otherwise get error for certain stocks
except Exception, e:
log.info('exception when ordering % ' % e)

Hi Attila,

You're running into a common problem presented (as is often the case) when there's an illiquid security or we don't have data for that security at that bar. You can fix it by using something like 'if stock in data: order_X' in your code or view here https://www.quantopian.com/help#api-common-errors for more information.

Thanks Seong.

Hi Seong,

Perhaps you could elaborate on what's going on in the back-end. My understanding is that once a given stock has its first historical trade of the backtest, forward-filling of price data is automatic (see https://www.quantopian.com/posts/thinly-traded-stocks-why-no-gaps). So, there will be no empty bars going forward from that first bar. But then, at some point, if the stock stops trading, do you stop forward filling? Or is the stock just dropped altogether from the data object? Or something else?

And what if I use 'history'? When does the filling of prices in data start? I realize that the DataFrame returned by history may or may not be forward-filled depending on the options; I'm wondering if the use of history will change the start date of the filling of prices in the object 'data'.

Grant