How to use filter() to remove all strings from a list with strings and numbers?

Question

I have a list for example

list = [1,2,3,'t',4,5,'fgt',6,7,'string']

and I want to use the filter() function to remove all the strings to leave just numbers. I can do it the normal method, but I cant do it with the filter method...any tips?

so:

list(filter(type(i)==str,a)))

wouldn't work...I tried to use it, but that still doesn't work:

Traceback (most recent call last):
  File "<pyshell#11>", line 1, in <module>
    list(filter(type(a[-1])==str,a))
TypeError: 'bool' object is not callable

Answer 1

While you could use filter for this, don't. You'd need a lambda function to do it, and it would be both slower and less readable than an equivalent list comprehension or generator expression. Instead, just use the listcomp or genexpr:

old_list = [1,2,3,'t',4,5,'fgt',6,7,'string']
new_list = [x for x in old_list if isinstance(x, (int, float))]
# or to remove str specifically, rather than preserve numeric:
new_list = [x for x in old_list if not isinstance(x, str)]

That's much more straightforward than the filter+lambda equivalent:

new_list = list(filter(lambda x: isinstance(x, (int, float)), old_list))

As noted in COLDSPEED's answer, to be generally accepting of all "number-alikes" you should actually use isinstance with numbers.Number; using (int, float) handles the literal types, but wouldn't handle complex, fractions.Fraction, decimal.Decimal, or third-party numeric types.

Answer 2

If you're looking for a filter, you can shape your lambda to be a bit more elegant.

from numbers import Number
new_list = list(filter(lambda x: isinstance(x, Number), old_list)) 

numbers.Number is an injected superclass of int and float, and complex. For real types, use numbers.Real instead.