grep - Regex match pattern with optional character and min/max length

Question

Match starts with X, which is followed by 3 to 6 capital words or numbers, must containt at least one letter and can have optional dash -. Example of matches:

XABC
XA-BC8D
X-ABC
XB72D-

Example of non-matches:

X123      //no letters
XAB       //too short
XABC-123  //too long
XA--BC    //too many -

I tried with X(?=.{3,6}$)[A-Z0-9]*-?[A-Z0-9]*, but it has many problems. It can match something like X---. Regex should be compatible with grep.

Answer 1

You may use this regex with gnu grep using -P (PCRE):

grep -P '^X(?!([^-]*-){2})(?=[^A-Z]*[A-Z])[A-Zd-]{3,6}$' file

XABC
XA-BC8D
X-ABC
XB72D-

RegEx Demo

RegEx Details:

• ^: Start
• X: Match letter X
• (?!([^-]*-){2}): Negative lookahead to assert that there are never more than one hyphens ahead
• (?=[^A-Z]*[A-Z]): Positive lookahead to assert presence of at least one uppercase letter ahead
• [A-Zd-]{3,6}: Match an uppercase letter or digit or - 3 to 6 times
• $: End

---

If you don't have gnu grep installed then you may consider this awk:

awk -F- 'NF<3 && /^X[A-Z0-9-]{3,6}$/ && /.[A-Z]/' file

XABC
XA-BC8D
X-ABC
XB72D-