In a regular expression, capture one thing, or another, or both

Question

Here is a piece of text I have :

I am trying to use regex to capture one group which is a alpha-numeric code (G1 or 27C) or another group which is a float (xx.xx), or both of them.

For this example, I want this specific return:

(G1,13.00)
(G1,3.00)
(,0.00)
(27C,)

this is the closest solution I have :

\(?:(\w+) +(\d+\.\d+))|(?: +(\d+\.\d+))|(?: +(\w+))\

The problem with this solution is that the last 2 rows values are captured in the 3rd and a 4th group:

(G1,13.00,,)
(G1,3.00,,)
(,,0.00,)
(,,,27C)

Any ideas on how to solve it ?

I have found this question which is close : In a regular expression, match one thing or another, or both but it answers how to match and not to capture.

With PyPi regex module, you could use `(?|(\w+) +(\d+\.\d+)| +(\d+\.\d+)| +(\w+))` — Wiktor Stribiżew, Jul 03 '17 at 20:43

logi-kal · Accepted Answer · 2017-07-03T14:37:47.583

1

You just need same patterns to be matched by same capturing groups.

This will return always 2 groups:

(\w+)? *\b(\d+\.\d+)?

EDIT: if you want no empty match, try this:

(?=\w)(\w+(?=$|\s))? *\b(\d+\.\d+)?

edited Jul 03 '17 at 14:37

answered Jul 03 '17 at 13:53

logi-kal

thanks but It matches with 0 characters and then I give me too much matches. – GregOizo Jul 03 '17 at 14:28
@GregOizo I edited the answer, now let me check if the regex is simplifiable – logi-kal Jul 03 '17 at 14:39
Well, I think it is ok – logi-kal Jul 03 '17 at 14:43

1 Answers1