Regex to match C-1234567-01-A3

Daravon · 07-23-2010, 09:11 AM

At work, I run the product scriber. The scriber scribes the product number on the product. A typical product number is

C-1234567-01-A3 (for example)

The C and the hyphens never change. So basically all valid scribes are of the form

C-*some seven numerical digits*-*some two numerical digits*-*some alphanumeric checksum*

I'm not a programmer but I need to ask the programmer to check the scribe's form. I don't want to actually read the scribe and see if it's correct, I just need to verify the syntax, so that we can raise a flag if there's no way it could be correct.

I want to communicate the requirements for a valid scribe syntax with a regex so that later there's no ambiguity about what I asked for. Problem is I don't know how to express it precisely. I originally came up with

C-[0-9]-[0-9]-[A-Z0-9]

This says what the 3 fields can contain, but it doesn't express the fact that the first number must be exactly 7 numeric digits, and the sequence number must be exactly 2, and the checksum must be exactly 2 alphanumeric digits.

How can I communicate these requirements unambiguously in a form that any programmer will understand?

ghostdog74 · 07-23-2010, 09:20 AM

Code:

$ var="test C-1234567-01-A3 test"
C-1234567-01-A3
$ echo $var | grep -oE "C-[0-9]{7}-..-.."
C-1234567-01-A3
$ var="test C-123456-01-A3 test"
$ echo $var | grep -oE "C-[0-9]{7}-..-.."

Daravon · 07-23-2010, 09:35 AM

I didn't expect an answer in the form of code; I'm afraid I don't speak bash. Maybe if it was perl...

I think I see what you did there partly though, so maybe what I'm looking for is

Code:

C-[0-9]{7}-[0-9]{2}-[A-Z0-9]{2}

?

pixellany · 07-23-2010, 09:41 AM

You are correct!!

The only other thing to note is that ghostdog put the -E flag in the grep statement. This means "use extended regular expression (ERE) syntax".

(I don't recall what about the expression is ERE, but you can read up on it in any good tutorial---eg: http://www.grymoire.com/Unix/Regular.html

grail · 07-23-2010, 10:01 AM

Quote:

I don't recall what about the expression is ERE

The curly bracketed grouping would need backslashes and stuff without the -E

Daravon · 07-23-2010, 10:07 AM

Cool, I created a text file containing a bunch of phony product numbers and one good product number, and now I'm using grep to test my regex. Surprisingly it pulled out a number I intended as a phony. I'm not sure it's a problem though, because checksums might not always contain letters. I'm not sure about that. I guess I need to know more about my checksum.

Code:

chaz@jarvis:~$ cat scribes.txt
C-1234567-01-A3 B-1234567-01-A3 C-1234567-A1-A3 C-123X567-01-A3 c-1234567-01-A3 C-1234567-01-03
chaz@jarvis:~$ cat scribes.txt | grep -oE "C-[0-9]{7}-[0-9]{2}-[A-Z0-9]{2}"
C-1234567-01-A3
C-1234567-01-03
chaz@jarvis:~$