How to test a regular expression with grep

A grep pattern that matches one good line can still catch the same text inside a bad line, especially when a script later treats every matching row as valid. Testing the regex against nearby failures first shows whether the pattern is only finding a substring or really matching the intended line shape.

grep prints whole input lines that contain a match by default. Anchors such as ^ and explicit separators decide whether a match can start anywhere in the line or must begin at the line boundary and stop before the next field.

grep -E uses extended regular expressions, which keep interval counts such as {4} readable. Plain grep uses basic regular expressions by default, so the same count needs escaped braces as \{4\}. The sample uses [0-9] for an ASCII ticket-code prefix; browser regex testers and PCRE examples can use different syntax, so the final proof should come from grep in the same shell environment where the pattern will run.

1. Create a small fixture with intended matches and near misses.

   tickets.txt

   OPS-2026 ready
   ops-2026 lower-case prefix
   OPS-20260 extra digit
   DBA-4112 approved
   WEB-2026 pending
   OPS-20A6 typo

   Keep at least one line that should match and several lines that look close but must stay out.

2. Run the first candidate pattern with grep -E.

   $ grep -E 'OPS-[0-9]{4}' tickets.txt
   OPS-2026 ready
   OPS-20260 extra digit

   The second line proves the pattern is too broad for a ticket prefix check because it finds OPS-2026 inside OPS-20260.

3. Add a start anchor and a field boundary to reject the near miss.

   $ grep -E '^OPS-[0-9]{4}[ ]' tickets.txt
   OPS-2026 ready

   ^ requires the match to start at the beginning of the line, [0-9]{4} requires exactly four digits, and [ ] requires a literal space after the code.

4. Use basic regular expression escaping when the command must run as plain grep.

   $ grep '^OPS-[0-9]\{4\}[ ]' tickets.txt
   OPS-2026 ready

   Use grep -E when the pattern needs unescaped ?, +, {}, |, or grouping characters. Use plain grep only when the deployed command expects basic regular expression notation.

5. Remove the temporary fixture after the pattern is confirmed.

   $ rm tickets.txt