Why is filter object not cast to str in python?

Question

Python

Actually, why do you have to fence this construction:

b = ''.join(list(filter(lambda c: c.islower() or c.isdigit(), a)))

instead of

b = str(filter(lambda c: c.islower() or c.isdigit(), a))

Well, after all, why would a non-lazy language suddenly need lazy default functions?
I'm not a holivar for the sake of it, I'm really interested in the reasons for this, in my opinion, not the most successful design.

Answer 1

Actually, why do you have to fence this construction:

And why are you bringing the filtering result to list()?

>>> b = ''.join(filter(None, 'abcde'))
>>> b
'abcde'
>>>

Answer 2

Why change anything at all?
>>> filter(lambda c: c.islower() or c.isdigit(), '[email protected]#')
'asd123'
The description of filter clearly states: If iterable is a string or a tuple, the result also has that type;
But this is if you have 'a' - a string. If 'a' is a list, then you first convert it to a string.
But this is true for the second python - in the third filter it is not a function, but an iterator that needs to be processed before you get a string:
>>> ''.join(filter(lambda c: c.islower() or c.isdigit(), '[email protected]#'))
'asd123'
>>>