Weeken with python (part 2)

In this serial, I write about what I learned about Python over the weekend. This week, I talk about 2 problems (1) function argument ’s default value and (2) function variable argument.

Function argument ’s default value

Python supports default value for function argument like C++. But function definition (using def keyword) actually a statement, so the default value of function argument will be calculated at the time function was defined. It means, when you set a default value for an argument, this value will change when you change it in the function body. Let ’s make an example:

import random

def make_array(data=[]):
    data.append(3)
    print data

The first time, you call make_array, data was set to new list object. The object is bind with data. Later, every time you call make_array, the object will change, so the size of data is increased with a number of the function call:

>>> make_array()
[3]
>>> make_array()
[3, 3]
>>> make_array()

The best practice to deal with this problem is we should set default value is None:

>>> def make_array(data=None):
        if not data:
            data = []
        data.append(3)
        print data
>>> make_array()
[3]
>>> make_array()
[3]
>>> make_array()
[3]
>>> make_array()
[3]

Variable arguments

Python supports variable argument and key-value variables arguments. Let ’s start with an example:

>>> def test_variable_args(*args):
        for arg in args:
          print arg
>>> test_variable_args(10, 20, 30, 40)
10
20
30
40

>>> def test_key_value_variable_arguments(**kvargs):
        for key, value in kvargs:
            print "%s: %s" % (key, value)
>>> test_key_value_variable_arguments({ "name" : "kiennt", "age" : 24 })
name: kiennt
age: 24