Parameters

Python offers flexible mechanisms for passing arguments to functions:

 1>>> x, y = 2, 3
 2>>> def func1(u, v, w):
 3...     value = u + 2 * v + w**2
 4...     if value > 0:
 5...         return u + 2 * v + w**2
 6...     else:
 7...         return 0
 8...
 9>>> func1(x, y, 2)
1012
11>>> func1(x, w=y, v=2)
1215
13>>> def func2(u, v=1, w=1):
14...     return u + 4 * v + w**2
15...
16>>> func2(5, w=6)
1745
18>>> def func3(u, v=1, w=1, *tup):
19...     print((u, v, w) + tup)
20...
21>>> func3(7)
22(7, 1, 1)
23>>> func3(1, 2, 3, 4, 5)
24(1, 2, 3, 4, 5)
25>>> def func4(u, v=1, w=1, **kwargs):
26...     print(u, v, w, kwargs)
27...
28>>> func4(1, 2, s=4, t=5, w=3)
291 2 3 {'s': 4, 't': 5}
Line 2

Functions are defined with the def statement.

Line 5

The return statement is used by a function to return a value. This value can be of any type. If no return statement is found, the value None is returned by Python.

Line 11

Function arguments can be entered either by position or by name (keyword). z and y are specified by name in our example.

Line 13

Function parameters can be defined with default values that will be used if a function call omits them.

Line 18

A special parameter can be defined that combines all additional positional arguments in a function call into one tuple.

Line 25

Similarly, a special parameter can be defined that summarises all additional keyword arguments in a function call in a dictionary.

Options for function parameters

Most functions need parameters. Python offers three options for defining function parameters.

Positional parameters

The simplest way to pass parameters to a function in Python is to pass them at the position. On the first line of the function, you specify the variable name for each parameter; when the function is called, the parameters used in the calling code are assigned to the function’s parameter variables based on their order. The following function calculates x as a power of y:

>>> def power(x, y):
...     p = 1
...     while y > 0:
...         p = p * x
...         y = y - 1
...     return p
...
>>> power(2, 5)
32

This method requires that the number of parameters used by the calling code exactly matches the number of parameters in the function definition; otherwise, a type error exception is thrown:

>>> power(2)
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
TypeError: power() missing 1 required positional argument: 'y'

Function parameters can have default values, which you can declare by assigning a default value in the first line of the function definition, like this:

def function_name(param1, param2=Standardwert2, param3=Standardwert3, ...)

Any number of parameters can be given default values, but parameters with default values must be defined as the last in the parameter list.

The following function also calculates x as a power of y. However, if y is not specified in a function call, the default value 5 is used:

>>> def power(x, y=5):
...     p = 1
...     while y > 0:
...         p = p * x
...         y = y - 1
...     return p
...

You can see the effect of the standard argument in the following example:

>>> power(3, 6)
729
>>> power(3)
243

Parameter names

You can also pass arguments to a function by using the name of the corresponding function parameter rather than its position. Similar to the previous example, you can enter the following:

>>> power(y=6, x=2)
64

Since the arguments for the power are named x and y in the last call, their order is irrelevant; the arguments are linked to the parameters of the same name in the definition of the power, and you get back 2^6. This type of argument passing is called keyword passing. Keyword passing can be very useful in combination with the default arguments of Python functions when you define functions with a large number of possible arguments, most of which have common default values.

Variable number of arguments

Python functions can also be defined to handle a variable number of arguments. This is possible in two ways. One method collects an unknown number of arguments in a list. The other method can collect an arbitrary number of arguments passed with a keyword that has no correspondingly named parameter in the function parameter list in a dict.

For an indeterminate number of positional arguments, prefixing the function’s final parameter name with a * causes all excess non-keyword arguments in a function call, that is, the positional arguments that are not assigned to any other parameter, to be collected and assigned as a tuple to the specified parameter. This is, for example, a simple way to implement a function that finds the mean in a list of numbers:

>>> def mean(*numbers):
...     if len(numbers) == 0:
...         return None
...     else:
...         m = sum(numbers) / len(numbers)
...     return m
...

Now you can test the behaviour of the function, for example with:

>>> mean(3, 5, 2, 4, 6)
4.0

Any number of keyword arguments can also be processed if the last parameter in the parameter list is prefixed with **. Then all arguments passed with a keyword are collected in a dict. The key for each entry in the dict is the keyword (parameter name) for the argument. The value of this entry is the argument itself. An argument passed by keyword is superfluous in this context if the keyword with which it was passed does not match one of the parameter names in the function definition, for example:

>>> def server(ip, port, **other):
...     print(
...         "ip: {0}, port: {1}, keys in 'other': {2}".format(
...             ip, port, list(other.keys())
...         )
...     )
...     total = 0
...     for k in other.keys():
...         total = total + other[k]
...     print("The sum of the other values is {0}".format(total))
...

Trying out this function shows that it can add the arguments passed under the keywords foo, bar and baz, even though foo, bar and baz are not parameter names in the function definition:

>>> server("127.0.0.1", port="8080", foo=3, bar=5, baz=2)
ip: 127.0.0.1, port: 8080, keys in 'other': ['foo', 'bar', 'baz']
The sum of the other values is 10

Mixing argument passing techniques

It is possible to use all the argument passing techniques of Python functions at the same time, although this can be confusing if you don’t do it carefully. Positional arguments should come first, then named arguments, followed by indefinite positional arguments with a simple *, and finally indefinite keyword arguments with **.

Mutable objects as arguments

Arguments are passed by object reference. The parameter becomes a new reference to the object. With immutable objects such as Tuples, Strings and Numbers, what is done with a parameter has no effect outside the function. However, if you pass a mutable object, such as a Lists, a Dictionaries or a class instance, any change to the object changes what the argument refers to outside the function. Reassigning the parameter has no effect on the argument.

>>> def my_func(n, l):
...     l.append(1)
...     n = n + 1
...
>>> x = 5
>>> y = [2, 4, 6]
>>> my_func(x, y)
>>> x, y
(5, [2, 4, 6, 1])

The variable x is not changed because it is unchangeable. Instead, the function parameter n is set so that it refers to the new value 6. However, there is a change in y because the list it refers to has been changed.

Checks

  • Write a function that can take any number of unnamed arguments and output their values in reverse order?