Is there a way to write better regex when we have so much ORs

Question

I have this following regex, starts with 09 and has a 2 digit number that should be only in some specific numbers, then 7 digits:

^[0][9](11|12|13|14|15|16|17|90|91|92|93|94|30|33|35|36|37|38|39|01|02|03|04|05|41|20|21|22|32|31|34|42)[0-9]{7}$

Or in python:

numbers = r"11|12|13|14|15|16|17|90|91|92|93|94|30|33|35|36|37|38|39|01|02|03|04|05|41|20|21|22|32|31|34|42"
phone_number_regex = rf"^[0][9]({numbers})[0-9]{7}$"

Is there a way to make this regex shorter? there is too much |s.

"there is too much `|` s" - and? Is it a problem? As an alternative, you could use regex to verify that there are two digits and check these digits against the allowed list in python, not regex. So you'd still have the list in the code, but the regex would be shorter. — Sergio Tulentsev, Jan 13 '22 at 15:08

JvdV · Accepted Answer · 2022-01-13T15:17:22.527

4

Looking at the alternatives, it seems the following matches the same things:

^09(0[1-5]|1[1-7]|2[012]|3[0-9]|4[12]|9[0-4])[0-9]{7}$

edited Jan 13 '22 at 15:17

answered Jan 13 '22 at 15:09

JvdV

70,606
8
39
70

If numbers in this list have business meaning (like allowed area codes in phone number), then compression like this completely hides it. Yes, the regex is shorter, but at what cost? :) – Sergio Tulentsev Jan 13 '22 at 15:12
1

Depending on your personal experience, this form of writing alternatives is perfectly readable @SergioTulentsev. Either way, the question was if it could be written "shorter". – JvdV Jan 13 '22 at 15:14
@JvdV: yeah, I'm just thinking out loud. Your version is not identical to the original expression, btw. It rejects 20, for example. Maybe others too. My point was: the list may be long, but it's nearly impossible to mess it up. Unlike this scheme. Especially when one needs to add/remove numbers. – Sergio Tulentsev Jan 13 '22 at 15:16
@SergioTulentsev, good point indeed. I missed the 20. Will edit now. – JvdV Jan 13 '22 at 15:16
1

Anyway, good stuff. +1 – Sergio Tulentsev Jan 13 '22 at 15:20
This is what I have used: `^09(0[1-5]|1[1-8]|2[0-2]|3[0-9]|4[12]|9[0-4])[0-9]{7}$`. Thank you, it helped so much – Mojtaba Arezoomand Jan 13 '22 at 15:25
1

You could also change the `[0-9]{7}` to `\d{7}`. – mrCarnivore Jan 13 '22 at 16:29

score 1 · Answer 2 · answered Jan 13 '22 at 15:10

1

I'm on mobile so I can't do the string split and sort to get the full thing, but what you can do is use character ranges to handle each group.

11|12|13|14|15|16|17|...

becomes

1[1-7]|...

and so on.

answered Jan 13 '22 at 15:10

samthecodingman

23,122
4
30
54

Thank you so much, it was exactly what I was looking for. – Mojtaba Arezoomand Jan 13 '22 at 15:11

Is there a way to write better regex when we have so much ORs

2 Answers2