Combination of map() and filter()

Question

I just had an idea that I find pretty intriguing: A combination between map() and filter() using generator as predicate and yield from. To make it short, here's the code:

def map_filter(function, iterable):
    """convert and filter a sequence"""
    for i in iterable:
        yield from function(i)

Now, what's the deal? Basically, this is a combination between the aforementioned two functions which incorporates both their functionality. Actually, the possibility to pass additional parameters to map() is still missing, though that's a minor detail IMHO and it could be extended for that. Here's a comparison, generating squares of numbers:

def function(x):
    return x * x
res = map(function, range(0, 10))
print(list(res))

def function(x):
    yield x * x
res = map_filter(function, range(0, 10))
print(list(res))

Here's another, filtering odd numbers:

def function(x):
    return x % 2 == 1
res = filter(function, range(0, 10))
print(list(res))

def function(x):
    if x % 2 == 1:
        yield x
res = map_filter(function, range(0, 10))
print(list(res))

And the last one, combining the two above:

def function1(x):
    return x * x
def function2(x):
    return x % 2 == 1
res = map(function1, filter(function2, range(0, 10)))
print(list(res))

def function(x):
    if x % 2 == 1:
        yield x * x
res = map_filter(function, range(0, 10))
print(list(res))

Notes and questions:

• First question of course, am I reinventing the wheel or would this be a useful addition to Python (e.g. in itertools)?
• Are there any problems or downsides to this? One I see is the slightly more complex interface, but it seems a useful tool for general sequence manipulation.
• map_filter() requires a generator (with yield), so I can't use lambda with it (Can I?). For very small filtering/mapping functions, the amount of code is larger therefore, but otherwise you can probably write more terse code using a combined filtering/mapping function.
• One of the best selling points is that not only can you use map_filter() to remove elements from a sequence, but you can also use it to inject additional elements.
• I'm of course open for any suggestions or improvements.

Answer 1

Use list comprehensions with conditional instead. For example:

squared_evens = [n*n for n in range(10) if n % 2 == 0]

squared_evens
>>> [0, 4, 16, 36, 64]

Answer 2

You can define generator expressions:

>>> values = range(0, 10)
>>> evens = (value for value in values if not value % 2)
>>> even_squares = (even * even for even in evens)
>>> list(even_squares)
[0, 4, 16, 36, 64]