Hello Arnaud,

If you add the line:

print stock.price  

you will see that there is no data for LEH after 2008-09-17. The 'record' feature continues to display the last values - I think this is the bug.

P.

I think that you're both right, actually.

record() does forward-fill values, and that is what you're seeing in the graph.

That said, we don't have a clean way to handle stocks that are de-listed. Sometimes they go bankrupt and go to zero, sometimes they go bankrupt and come back at a lower value, sometimes they get de-listed and start trading OTC, etc. We don't have a canonical datasource that helps us re-value those stocks that are no longer trading. It's something we can improve in the future.

Hi Dan,

Is this planned to be fixed anytime soon?
...in the meantime, is there a workaround to detect that a stock is de-listed?

@Peter:
Thanks for the answer. Interestingly, this seems related to another bug:

If you only register a single stock, the "handle_data" is apparently not even called after the stock is de-listed.

def initialize(context):  
    context.stock = sid(11326) # LEH  
    # context.aapl = sid(32146) # AAPL  
    context.count = 0  
def handle_data(context, data):  
    context.count += 1  
    record(count = context.count)  

In the above code, you will see two different curves depending if only LEH, or both LEH,AAPL are included. In the first case, the counter will stop increasing once LEH is de-listed. In the second case, the counter will continue increasing normally. Surprising, isn't it? Just try it out ;)

Hello Arnaud,

This is the converse example where 'record' would give an error if the security has not yet started trading so a check is required. With start of 2010-01-01 run:

def initialize(context):  
    pass

def handle_data(context, data):  
    if sid(40207) in data:  
        record(CELG_Z=data[sid(40207)].close_price)  

To check if current data exists for a security try

if data[sid(11326)].datetime < get_datetime():  

P.

Thanks for the tip, the datetime check works!