Function for assigning variables
 

Here's something that I've wished for a number of times so far: the ability to create a function (taking X) that creates X variables, even if they are all initialized to the same value. For example, f(3) would assign var1 = 0, var2 = 0, and var3 = 0.

My program runs an experiment, and I'd love for the number of sections to be a variable that I could change at will (so I could expand it later if I wanted). Each section requires a certain number of variables for keeping track of things like how many problems the participant solved correctly. So you can see how useful the above would be to me...

Anyway, I assume it can't be done. But are there other ways to deal with my problem? I'd love to find out.

Re: Function for assigning variables
 

If I understand you correctly, there are several ways:

# Update the local namespace
>>> for i in range(25):
... exec "%s = %d" % ('var'+str(i), 0) in None

# Update the global namespace
>>> dd = {}
>>> for i in range(25):
... dd['var'+str(i)] = 0
>>> globals.update(dd)

# Update the local namespace with dictionary dd
>>> for key in dd:
... exec "%s = %s" % (key, repr(dd[key])) in None
...
>>>

# Update a module namespace with dd
>>> import P3D
>>> P3D.__dict__.update(dd)

>>> var0
0
>>> var1
0
>>> var24
0
>>>

Re: Function for assigning variables
 

hi.. I assume you can better handle this with using class. So I have a small piece of code, I assume i understand your problem.

class Exp:

    """This is an experiment class"""

    def __init__(self, section, n):

        temp = dict([i+1, 0] for i in range(n))

        self.block = {section: temp}



    def func(self, section):

        for k, v in self.block[section].iteritems():

            print k, v

            print self.block[section][k]



if __name__ == '__main__':

    obj = Exp('Qsection', 3)

    print obj.func('Qsection')

lets say, you are conducting an experiment and it has several blocks, and each block has sections. And you want process the section's data.
The section's inputs vary. So, I created a class, which includes a dictionary(a hash).

The structure of the hash is
block => [section_name] => [v1:value, v2:value...]

so you can dynamically add number of keys to hash which is equal to creating that many number of variables.

Add functions to class as per your requirement.


Hope this helps.




kath.

Re: Function for assigning variables
 

Hey ... weird... I thought I had posted here earlier. Anyways, assuming that you are representing each section with a single integer, you might do well to simply maintain a list of sessions. In that case, your function would be

def f(x):

   return [0]*x

If it turns out you need more info for your sections, you could create a Section() class:

class Section(object):

   def __init__(self):

       self.score = 0

       self.name = ""

       ... # etc.



def f(x):

    return [Section() for i in range(x)]

Now, f(x) returns x number of Section objects ready for use.

Jeff

Re: Function for assigning variables
 

hey, I din't aware of your post.


kath.

Re: Function for assigning variables
 

Oh, sorry ... wasn't complaining about your post, which was great. I was just surprised to find myself reading the same question that I thought I'd replied to earlier... :lol:

Re: Function for assigning variables
 

Wow, a big thanks to all of you! Now to compile this information in the ol' brain and see if I can improve my experiment program...

I still honestly have a lot to learn about structuring my whole program. Right now I have a couple of different classes based on the type of page that the participant will be viewing (an instructions page, a trial page, or a questionnaire page), and I've got a multiway branch set up that directs them to the appropriate type based on what section of the experiment they're in... and the section is just an int that increments as they progress along. I suppose it's not too bad overall, for a first try anyway.

Anyway, my plan right now is to use solsteel's suggestion of the exec function in order to assign some features to the different trial blocks. We'll see how well it goes.