Python Lambda

Posted Thursday, 20 August 2009 at 16:44 by Andrew Liu
Tagged: python | web development
Read more blogs...

Using the lambda function in Python can be a little daunting, and I've always strayed away from it in the past.  It is very useful in doing calculations on the fly, and some applications of the Python lambda function can include reporting calculations, strenuous accounting functions, and number intensive programs.  Fortunately for me, I've kept away from those sorts of applications, and lambda is something that I know of, have used on the odd occasion, but never really fully understood it.

Today, I needed to use it, so I thought I'd dig up some of the documentation I've read on it in the past, go over it in more detail, and finally understand it enough so I can say that I've mastered it.

In short, lambda allows you to create an on-the-fly function.  Think of a once-off function - something that you would use briefly, then disregard.  Its similar to using that "i" variable in a loop; something that you simply need quickly to serve for a short purpose, then your program moves on.

In that sense, these extracts of Python code serve exactly the same purpose:

>>> def f (x): return x**2
...
>>> print f(8)
64
>>>
>>> g = lambda x: x**2
>>>
>>> print g(8)
64

Both extracts of Python code serve to produce the square of the given input parameter.  So whats the difference between these?

Firstly, the lambda function has no return function.  It always contains an expression which is always returned. 

Secondly, the lambda function is completed in a single line, whereas the function definition can span over multiple lines.  You cannot make a lambda function span multiple lines; it must be a single pythonic expression.  The result of that expression is the returned value for the lambda function.

To clear things up, lambda will return you a function, as show here:

>>> type(f)
<type 'function'>
>>> type(g)
<type 'function'>
>>>

Thus, the resulting function can be used anywhere else a function is expected, and will work in exactly the same way.

So where does this come in handy?  In my honest opinion, it isn't really that handy, because you can define a function which does the same thing, albeit in a few more lines of code.  I would think some Python purists would argue that the lambda function is "more correct" in some instances, and I would agree with this as well.  It can be very useful in streamlining Python code.  Perhaps the best example of this is when you look at the following problem, and the solutions I produce.

Problem

I have given sentence of varying length, and I would like to know how long each word is in that sentence.

Solution

Using "classical" Python (that is, using no lambda), this can be easily achieve in a series of loops as follows:

>>> a = "I have given sentence of varying length, and I would like to know how long each word is in that sentence."
>>> words = a.split(" ")
>>> lengths = []
>>> for word in words:
...     lengths.append(len(word))
...
>>> lengths
[1, 4, 5, 8, 2, 7, 7, 3, 1, 5, 4, 2, 4, 3, 4, 4, 4, 2, 2, 4, 9]
>>>

I could streamline this a bit with some smart do-it-one-line python, as follows:

>>> a = "I have given sentence of varying length, and I would like to know how long each word is in that sentence."
>>> words = a.split(" ")
>>> lengths = [ len(word) for word in words ]
>>> lengths
[1, 4, 5, 8, 2, 7, 7, 3, 1, 5, 4, 2, 4, 3, 4, 4, 4, 2, 2, 4, 9]
>>>

I can also achieve the same result using a lambda function, as follows:

>>> a = "I have given sentence of varying length, and I would like to know how long each word is in that sentence."
>>> words = a.split(" ")
>>> lengths = map(lambda x: len(x), words)
>>> lengths
[1, 4, 5, 8, 2, 7, 7, 3, 1, 5, 4, 2, 4, 3, 4, 4, 4, 2, 2, 4, 9]
>>>

In essence, all three solutions do the same thing, but hopefully you can see the benefit in using one of the last 2 solutions.  I personally favour the second solution, instead of using lambda, because it is a combination of "for" and "lists" that keep the code close to heart, instead of using a lambda function.  But there are cases when you must use a function (eg. when you actually pass functions around), and in this case, a lambda is the better solution all round.

How Can WebTop Help?

Learn how people like you are using WebTop to solve their database needs

Watch a video demonstation of CASE, our flagship web management system