Python 3 Pandas: Custom sorting a list of strings

I need to sort a list of strings such that "non-numeric" values are sorted alphabetically first, followed by "numeric" values, in ascending order.

For example, suppose the list is:

l = ['2.9', '-1', '0', 'ab2', 'xyz', 'ab']

I'd like the output to be:

sorted_l = ['ab', 'ab2', 'xyz', '-1', '0', '2.9']

So far, I can only do this if the strings are all "integers":

import functools

l=['1','0','-1','2', '-9']

def compare(x, y):
    return int(x) - int(y)

sorted_l = sorted(l, key = functools.cmp_to_key(compare))

It'd be even more ideal if I can do it without using functools.
Thanks

A possible solution:

def is_numeric(s):
  try:
    float(s)
    return True
  except ValueError:
    return False

l = ['2.9', '10', '0', 'ab2', 'xyz', 'ab']
l.sort(key=lambda x: (is_numeric(x), float(x) if is_numeric(x) else x))

Alternatively and more elegantly, we can follow @QuangHoang's suggestion:

def is_numeric(s):
  try:
    float(s)
    return True, float(s)
  except ValueError:
    return False, s

l = ['2.9', '100', '0', 'ab2', 'xyz', 'ab']
l.sort(key=lambda x: is_numeric(x))

Output:

['ab', 'ab2', 'xyz', '-1', '0', '2.9']