Linux Command Reference ยท Concepts
Globs vs regular expressions
Shell globs and regular expressions look alike but are different languages: compare their syntax, who interprets them, and where each is used on Linux.
Linux has two pattern languages that share some characters and almost nothing else. A glob (wildcard pattern) matches file names and is expanded by the shell: *.txt. A regular expression (regex) matches text and is interpreted by a program such as grep: ^[a-z]+\.txt$.
Mixing them up is one of the most common beginner mistakes: grep *.txt does not mean "search in txt files for...", and ls .* does not mean "everything starting with a dot".
ls *.txt # glob: the shell expands it grep 'error.*' log # regex: grep interprets it, so quote it
Try it
Press Run (or Ctrl + Enter) to see the result here.
Program input ยท what you'd type while the program runs
Type the answers before you press Run, one per line. Each time the program reads a line (with read or cat), it gets the next line from this box.
Press Run (or Ctrl + Enter) to see the result here.
Program input ยท what you'd type while the program runs
Type the answers before you press Run, one per line. Each time the program reads a line (with read or cat), it gets the next line from this box.
Comparison
| Meaning | Glob | Regex (ERE) |
|---|---|---|
| Any single character | ? |
. |
| Any run of characters | * |
.* |
| One of a set | [abc], [a-z] |
[abc], [a-z] |
| Not in a set | [!abc] (also [^abc] in Bash) |
[^abc] |
| Alternatives | {a,b} (Bash brace expansion, not a glob) or @(a|b) with extglob |
a|b |
| Zero or more of the previous item | *(x) with extglob |
x* |
| One or more | +(x) with extglob |
x+ |
| Optional | ?(x) with extglob |
x? |
| Anchors | implicit: whole name must match | ^, $ (otherwise matches anywhere) |
| Literal dot | . is literal |
\. |
| Hidden files | * skips a leading dot |
no such rule |
| Path separator | * and ? never match / |
. matches / |
The deepest difference is anchoring. A glob always matches the whole name, so *.txt needs the name to end in .txt. A regex finds a match anywhere in the text unless you add ^ and $.
Where each one is used
- Globs: shell command lines,
casepatterns,[[ x == pattern ]],find -name,-path,ls,rsync --include,.gitignore,${var#pattern}and${var//pattern/x}trimming and replacement. - Regexes:
grep,sed,awk,[[ x =~ re ]],find -regex,lessandvimsearches,rename(Perl),pgrep. - Hybrid:
gitpathspecs andtar --excludeaccept globs;ls -Itakes a glob;egreptakes a regex.
Usage notes
- The shell expands unquoted globs first. If the pattern matches files, the program sees the file list, not the pattern. If nothing matches, Bash passes the pattern through unchanged (unless
nullgloborfailglobis set). - Quote regexes so the shell does not touch them:
grep '^a.*z$' file. A pattern with$or*unquoted is a bug waiting to happen. - Dotfiles:
*never matches names starting with.unlessshopt -s dotglobis set;.*also matches.and..in some shells. - No recursion in plain globs.
**matches across directories only withshopt -s globstar. - Case: both are case-sensitive by default.
shopt -s nocaseglobandgrep -ichange it. - Extended globs (
shopt -s extglob) add?(...),*(...),+(...),@(...),!(...), which mimic a few regex operators but remain glob rules. - Performance: globbing happens before the command runs and is limited by the argument-length limit (
Argument list too long);findandxargsavoid that.