Use a negative lookahead and a negative lookbehind:
> s = "one two 3.4 5,6 seven.eight nine,ten"
> parts = re.split('\s|(?<!\d)[,.](?!\d)', s)
['one', 'two', '3.4', '5,6', 'seven', 'eight', 'nine', 'ten']
In other words, you always split by \s
(whitespace), and only split by commas and periods if they are not followed (?!\d)
or preceded (?<!\d)
by a digit.
DEMO.
EDIT: As per @verdesmarald comment, you may want to use the following instead:
> s = "one two 3.4 5,6 seven.eight nine,ten,1.2,a,5"
> print re.split('\s|(?<!\d)[,.]|[,.](?!\d)', s)
['one', 'two', '3.4', '5,6', 'seven', 'eight', 'nine', 'ten', '1.2', 'a', '5']
This will split "1.2,a,5"
into ["1.2", "a", "5"]
.
DEMO.