Hi Leon,

Assuming you want to call the function on the first day of trading, I suggest setting a boolean to check if the function has been called once already, like in this snippet. Let me know if I can help further.

def initialize(context):  
    #create a boolean value:  
    context.function_bool = True

    schedule_function(func=function,  
                      date_rule=date_rules.every_day(),  
                      time_rule=time_rules.market_open())  


def function(context, data):

    if context.function_bool == True:  
            #function logic goes here  
            context.function_bool == False

Note that if the function doesn't need to be called during the trading day, this may be a better option:

def initialize(context):  
    #create a boolean value:  
    context.function_bool = True

def before_trading_start(context, data):

    if context.function_bool == True:  
            function(context,data)  
            context.function_bool == False

def function(context,data)  
    #function logic goes here  

Also, even for calling the function within the trading day, I wouldn't necessarily embed the execution control as illustrated. For example, consider:

def initialize(context):  
    #create a boolean value:  
    context.function_bool = True

    schedule_function(func=function,  
                      date_rule=date_rules.every_day(),  
                      time_rule=time_rules.market_open())  


def function(context, data):

    if context.function_bool == True:  
            my_function(context,data)  
            context.function_bool == False

def my_function(context,data)  
    #function logic goes here  

This way, the actual function that does the computing is separated from the function call control (I ain't no computer programmer, but it seems that one would want to avoid the "don't do anything" option within a function, and not call the function in the first place).

Great tip. Thank you all =)

Another consideration would be once the function is called to delete it and free up the resources. I don't know how to do this in Python and the Quantopian API but it would seem to be a best practice from a coding standpoint. Perhaps someone can provide guidance?

Just used this, thanks!

note in Grant's code there was a small typo:

def initialize(context):  
    #create a boolean value:  
    context.function_bool = True

def before_trading_start(context, data):

    if context.function_bool == True:  
            function(context,data)  
            context.function_bool = False  #not ==

def function(context,data) :   #was missing :  
    #function logic goes here