How can I split this string up?

Question

I'm trying to split strings up preceding the places where there is a whole, 2 digit number surrounded by whitespace. Eventually I'd like this to work in Python, but I've been workbenching with sed and I can't figure it out.

My test data looks like this:

13 13 13 13 13 9:07.18 9:12.09 9:15.65
14 14 14 2:04.86 2:05.99 2:06.87 14 4:21.51 4:23.51 4:25.00 14 8:56.28 9:01.09 9:04.58
15 15 57.18 57.61 57.95 15 2:02.61 2:03.72 2:04.58 15 4:17.31 4:19.28 4:20.75 15 8:47.15 8:51.87 8:55.30
16 16 56.34 56.76 57.09 16 2:00.69 2:01.78 2:02.63 16 4:13.75 4:15.69 4:17.14 16 8:39.71 8:44.37 8:47.75
17 25.69 25.85 25.99 17 55.62 56.03 56.36 17 1:59.07 2:00.15 2:00.99 17 4:10.76 4:12.69 4:14.11 17 8:33.73 8:38.34 8:41.68
18 25.43 25.59 25.73 18 55.01 55.42 55.74 18 1:57.74 1:58.81 1:59.63 18 4:08.34 4:10.24 4:11.66 18 8:33.73 8:37.04
19 25.20 25.36 25.49 19 54.50 54.91 55.23 19 1:57.74 1:58.56 19 4:08.34 4:09.74 19 8:33.73

And I would like it to be split up like this (note the location of the commas ','):

13, 13, 13, 13, 13 9:07.18 9:12.09 9:15.65
14, 14, 14 2:04.86 2:05.99 2:06.87, 14 4:21.51 4:23.51 4:25.00, 14 8:56.28 9:01.09 9:04.58
15, 15 57.18 57.61 57.95, 15 2:02.61 2:03.72 2:04.58, 15 4:17.31 4:19.28 4:20.75, 15 8:47.15 8:51.87 8:55.30
16, 16 56.34 56.76 57.09, 16 2:00.69 2:01.78 2:02.63, 16 4:13.75 4:15.69 4:17.14, 16 8:39.71 8:44.37 8:47.75
17 25.69 25.85 25.99, 17 55.62 56.03 56.36, 17 1:59.07 2:00.15 2:00.99, 17 4:10.76 4:12.69 4:14.11, 17 8:33.73 8:38.34 8:41.68
18 25.43 25.59 25.73, 18 55.01 55.42 55.74, 18 1:57.74 1:58.81 1:59.63, 18 4:08.34 4:10.24 4:11.66, 18 8:33.73 8:37.04
19 25.20 25.36 25.49, 19 54.50 54.91 55.23, 19 1:57.74 1:58.56, 19 4:08.34 4:09.74, 19 8:33.73

The data above, is fairly regular in that the two digit whole numbers are in the range [13,19], but the range I should expect is [10,99].

Could somebody suggest a method to perform the above transformation? I've been at this with regex for a while now but I can't cover all the cases.

Answer 1

The look-ahead assertion (?=...) can solve this:

>>> a = """13 13 13 13 13 9:07.18 9:12.09 9:15.65
14 14 14 2:04.86 2:05.99 2:06.87 14 4:21.51 4:23.51 4:25.00 14 8:56.28 9:01.09 9:04.58
15 15 57.18 57.61 57.95 15 2:02.61 2:03.72 2:04.58 15 4:17.31 4:19.28 4:20.75 15 8:47.15 8:51.87 8:55.30
16 16 56.34 56.76 57.09 16 2:00.69 2:01.78 2:02.63 16 4:13.75 4:15.69 4:17.14 16 8:39.71 8:44.37 8:47.75
17 25.69 25.85 25.99 17 55.62 56.03 56.36 17 1:59.07 2:00.15 2:00.99 17 4:10.76 4:12.69 4:14.11 17 8:33.73 8:38.34 8:41.68
18 25.43 25.59 25.73 18 55.01 55.42 55.74 18 1:57.74 1:58.81 1:59.63 18 4:08.34 4:10.24 4:11.66 18 8:33.73 8:37.04
19 25.20 25.36 25.49 19 54.50 54.91 55.23 19 1:57.74 1:58.56 19 4:08.34 4:09.74 19 8:33.73"""

>>> print(re.sub("(\d{2}) (?=\d{2}( |$))","\g<1>, ", a))
13, 13, 13, 13, 13 9:07.18 9:12.09 9:15.65
14, 14, 14 2:04.86 2:05.99 2:06.87, 14 4:21.51 4:23.51 4:25.00, 14 8:56.28 9:01.09 9:04.58
15, 15 57.18 57.61 57.95, 15 2:02.61 2:03.72 2:04.58, 15 4:17.31 4:19.28 4:20.75, 15 8:47.15 8:51.87 8:55.30
16, 16 56.34 56.76 57.09, 16 2:00.69 2:01.78 2:02.63, 16 4:13.75 4:15.69 4:17.14, 16 8:39.71 8:44.37 8:47.75
17 25.69 25.85 25.99, 17 55.62 56.03 56.36, 17 1:59.07 2:00.15 2:00.99, 17 4:10.76 4:12.69 4:14.11, 17 8:33.73 8:38.34 8:41.68
18 25.43 25.59 25.73, 18 55.01 55.42 55.74, 18 1:57.74 1:58.81 1:59.63, 18 4:08.34 4:10.24 4:11.66, 18 8:33.73 8:37.04
19 25.20 25.36 25.49, 19 54.50 54.91 55.23, 19 1:57.74 1:58.56, 19 4:08.34 4:09.74, 19 8:33.73

So, the reg exp. you need is(\d{2}) (?=\d{2}( |$)) which means:

1. (\d{2}) => Store 2 numbers in group 1 and match an extra space.
2. (?=\d{2}( |$)) => match 2 numbers and 1 space or EOL, but don't consume them.

The key here is that by not consuming the second matched group, it will be processed again next time that the sub function is applied. Finally, \g<1>, will replace 1. with the same numbers and the additional ,.