In Python, lambda is just a way of defining functions inline,
a = lambda x: x + 1
print a(1)
and..
def a(x): return x + 1
print a(1)
..are the exact same.
There is nothing you can do with lambda which you cannot do with a regular function—in Python functions are an object just like anything else, and lambdas simply define a function:
>>> a = lambda x: x + 1
>>> type(a)
I honestly think the lambda keyword is redundant in Python—I have never had the need to use them (or seen one used where a regular function, a list-comprehension or one of the many builtin functions could have been better used instead)
For a completely random example, from the article "Python’s lambda is broken!":
To see how lambda is broken, try generating a list of functions fs=[f0,...,f9] where fi(n)=i+n. First attempt:
>>> fs = [(lambda n: i + n) for i in range(10)]
>>> fs[3](4)
13
I would argue, even if that did work, it's horribly and "unpythonic", the same functionality could be written in countless other ways, for example:
>>> n = 4
>>> [i + n for i in range(10)]
[4, 5, 6, 7, 8, 9, 10, 11, 12, 13]
Yes, it's not the same, but I have never seen a cause where generating a group of lambda functions in a list has been required. It might make sense in other languages, but Python is not Haskell (or Lisp, or ...)
Please note that we can use lambda and still achieve the desired
results in this way :
>>> fs = [(lambda n,i=i: i + n) for i in range(10)]
>>> fs[3](4)
7
Edit:
There are a few cases where lambda is useful, for example it's often convenient when connecting up signals in PyQt applications, like this:
w = PyQt4.QtGui.QLineEdit()
w.textChanged.connect(lambda event: dothing())
Just doing w.textChanged.connect(dothing) would call the dothing method with an extra event argument and cause an error. Using the lambda means we can tidily drop the argument without having to define a wrapping function.
