functional programming with python

To work with lists many c/c++ programmer use a simple for-loop. I will try to explain how this can be made a little bit easier. For that I use the self-explaining script language python.

Here is a simple example to filter all odd integers from a list in simple python syntax.

def filter_odd(myList): result = [] # initialize the result list  for num in myList: # iterate over the list  if num % 2 == 1: # modulo 2 return 1 for odd integers result.append(num) # add to the result return result 

The function call will return a list containing only odd integers (you can use the result of the modulo operation directly without comparing with 1). Try it directly on a interactive python console (as an applet on secapps.com).

>> print filter_odd([1,2,3,4,5]) [1, 3, 5]

try it functional

The function filter_odd could be a more generic method, if the condition could be a variable or used as a parameter:

def condition_odd(num): # the condition as separate function return num % 2 == 1 #condition_odd   def generic_filter(condition, myList): result = [] for num in myList: if condition(num): # the same condition like the first example, only as parameter result.append(num) return result #generic_filter

We can define a condition as a function and use it as a parameter in the generic_filter implementation to filter the values of list myList. Ok, you don't have to define a generic filter function, python offers the filter function in the standard language.

>> print filter_generic(condition_odd, [1,2,3,4,5]) [1, 3, 5] >>print filter(condition_odd, [1,2,3,4,5]) [1, 3, 5]

The new implementation does not say what to do, it say how to use the condition on a list and has much less code compared to first example.

Passing a function as a parameter is a key aspect of functional programming. But the condition_odd function will only be used for the line with the filter call. The next step will show you, how to use this function parameter without defining it separately.

>> print filter(lambda x: x % 2 == 1, [1,2,3,4,5]) [1, 3, 5]

The first parameter is the lambda construct: Create a temporary, anonymous function with x as parameter and return the result of the term "x % 2 == 1". I don't have to define a function and I can use the filter in one line. Save namespace with one-time functions and code lines.

list comprehension

One-line solutions are fine to write short code, but it also has to be readable. If the reader doesn't know what filter and lambda means, it can be confusing.

An other alternative for the one-line filter solution, and more confusing in python, is list comprehension. You can write the above example:

>> print [x for x in [1,2,3,4,5] if x % 2 == 1] [1, 3, 5]

The result contains items from the constant list, filtered with the term in the if-condition. [Updated] You have to use the exact syntax, not the simple variant like filter.

more functional helper: map and reduce

There are some other list functions. First the map function with calls for every item in the list a function. In a example we will double every integer from the list. This could also be written as a list comprehension.

>>print map(lambda x: x*2, [1,2,3,4,5]) [2, 4, 6, 8, 10] >>print [x**2 for x in [1,2,3,4,5]] [2, 4, 6, 8, 10]

Second function is reduce. It get a function and a list as parameter and will call the function with the first two items from the list. Then the result and the next item will be computed and so on. A simple exapmle will calculate the sum of a number list:

>>print reduce(lambda a, b: a + b, [1,2,3,4,5]) 15

This simple example can also resolved with the build-in sum function.

two examples

Now we will try some real world examples - project eulers first problem: Find the sum of all the multiples of 3 or 5 below 1000. Simply to resolve with a for-loop:

1. the condition term is x%3==0 or x%5==0
2. filter the values beetween 0 and 1000
3. calculate the sum

def eulerOne(myList): mySum = 0 for num in range(1000): if num%3==0 or num%5==0: mySum += num return mySum

But we want to do it with an one-liner, with functional programming or a list comprehension:

>>print reduce(lambda a,b:a + b, [x for x in range(1000) if x % 3 == 0 or x % 5 == 0]) 233168 >>print sum([x for x in range(1000) if not (x%3 and x%5)]) 233168

The second line show a shorter version with sum and a little bit boolean algebra but it will not so clear readable.

As a second example filter all persons who older than 18:

>>persons = [{'age': 5, 'name': 'peter'}, {'age': 20, 'name': 'paul'}] >>print [x['name'] for x in persons if x['age'] > 18] ['paul'] >>print map(lambda x:x['name'], filter(lambda x: x['age']>19, persons)) ['paul']

summary

Here the summary with the python doc-strings to the new learned functions:

def name(arguments): return expression

lambda arguments: expression

[Updated]: filter and map are still in python 3000 in the iterator variant, the reduce-function is moved in the functools module.