Remove empty strings from tuples inside a list

list, python, regex, tuples

Solution

A nested list comprehension with a tuple constructor:

>>> lst = [('', '', '20', '500'), ('21', 'failed', '', '')]
>>> [(tuple(int(x) if x.isdigit() else x for x in _ if x)) for _ in lst]
[(20, 500), (21, 'failed')]

For each tuple (`_`) in `lst`, construct a `tuple` with a generator expression. The tuple constructor alone is below:

tuple(int(x) if x.isdigit() else x for x in _ if x)

Which seems confusing, but I'll break it down. For each string `x` in tuple `_` (which is a tuple in `lst`), construct a tuple. `if x` checks to see if the string is not empty. (If string `x` is empty, `x` is false.) `if x`, the generator expression will produce either `x` or `int(x)` depending on whether `x` in question is a digit in string form. (Trying to turn a non-numeric string into an integer will result in an exception.)

For each tuple `_` in `lst`, the generator creates a new, identical tuple, except the empty, false strings are filtered out and any digit strings are converted into `int`s.

Above code is equivalent to:

new_lst = []

for _ in lst: # For each tuple in lst
    for x in _: # For each string in tuple
        temp_tuple = ()
        if x: # Only add to tuple if string is not empty
            if x.isdigit(): # If x is a digit in string form
                temp_tuple += (int(x),) # Convert to int
            else:
                temp_tuple += (x,) # Keep string
    new_lst.append(temp_tuple)

Problem

Right now I have three lists that are generated from a RE findall function, and I am trying to remove some of the empty strings from the tuples inside the list. And the numbers should be converted to integer in the process too: Got:`[('', '', '1', '1')]` Expected: `[(1, 1)]` Got: `[('', '', '20', '500'), ('21', 'failed', '', '')]` Expected: `[(20, 500), (21, 'failed')]` Got: `[('3', 'failed', '', ''), ('', '', '48', '23'), ('', '', '96', '0')]` expected: `[(3, 'failed'), (48, 23), (96, 0)]` Any ideas?

Original source