Question

I’m doing a database insert script in pycassa. I want to set up a public static class that defines some variables that will get used a lot by other functions later on. Heres what I have…

class ks_refs():
    pool = ConnectionPool('TweetsKS')

    user_name_cf = self.cf_connect('UserName')
    user_tweet_cf = self.cf_connect('UserTweet')

    def cf_connect(column_family):
        cf = pycassa.ColumnFamily(self.pool, column_family)
        return cf

I haven’t even tried to run this yet because I’m sure it wont work. You can see I want this static variable ‘pool’ first, and then set up user_name_cf and user_tweet_cf (and some more later) using the cf_connect method which needs ‘pool’ to work.

I know I could put that method outside the class, or I could have this non-static and make an instance of it, but I want to try this because this is what I really want (before I was just using globals but I think a static class holding all this is the best idea)

Answer 1

I think you want to have a class method instead:

@classmethod
def cf_connect(cls, column_family):
    cf = pycassa.ColumnFamily(cls.pool, column_family)
    return cf

Now you can refer to the pool defined on your class with ease.

Your user_name_cf and user_tweet_cf ‘attributes’ will not work, however. You can add these after having created the class definition:

class ks_refs():
    pool = ConnectionPool('TweetsKS')

    @classmethod
    def cf_connect(cls, column_family):
        cf = pycassa.ColumnFamily(cls.pool, column_family)
        return cf

user_name_cf = ks_refs.cf_connect('UserName')
user_tweet_cf = ks_refs.cf_connect('UserTweet')

where they are then module-level constants, or you can add them to the class as attributes after the fact:

ks_refs.user_name_cf = ks_refs.cf_connect('UserName')
ks_refs.user_tweet_cf = ks_refs.cf_connect('UserTweet')