The challenge

Write your own implementation of the built-in split function in Python.

The following rules must be adhered to:

  • the function cannot use, in any way, the original split or rsplit functions,
  • the new function must be a generator,
  • it should behave as the built-in split, so it will be tested that way — think of split() and split('')

The solution in Python

Option 1:

 1
 2
 3
 4
 5
 6
 7
 8
 9
10

import re
def my_very_own_split(string, delimiter = None):
    if delimiter == '': raise ValueError('empty delimiter')
    if delimiter == None: delimiter = '\s+'
    else: delimiter = re.escape(delimiter)
    pos = 0
    for m in re.finditer(delimiter, string):
        yield string[pos:m.start()]
        pos = m.end()
    yield string[pos:]

Option 2:

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11

import re

def my_very_own_split(s, dell = None):
    f = []
    if dell=='': raise SyntaxError
    if dell:
        while dell in s:
            f.append(s[:s.find(dell)])
            s = s[s.find(dell)+len(dell):]
        return f+[s]
    return re.split(r'[\s]*',s)

Option 3:

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12

from string import whitespace, maketrans
convert_whitespace = lambda s: s.translate(maketrans(whitespace, ' ' * len(whitespace)))

def my_very_own_split(strng, sep=None):
    start, fil = 0, False
    if sep == '': raise ValueError("empty separator")
    if sep is None: sep, fil, strng = ' ', True, convert_whitespace(strng)
    while start <= len(strng):
        end = strng.find(sep, start)
        if end < 0: end = len(strng)
        if not fil or strng[start: end]: yield strng[start: end]
        start = end + len(sep)

Test cases to validate our solution

1
2
3
4
5
6
7
8

s, d = 'abc,def,ghi', ','
test.assert_equals(list(my_very_own_split(s, d)), ['abc', 'def', 'ghi'])

s, d = 'This is a test', ' '
test.assert_equals(list(my_very_own_split(s, d)), ['This', 'is', 'a', 'test'])

s, d = 'This is a test', ','
test.assert_equals(list(my_very_own_split(s, d)), ['This is a test'])