Linux Command Reference ยท Bash Syntax
Regex matching in Bash (=~)
Reference for regular expression matching in Bash: the =~ operator inside [[ ]], extended regex syntax, capture groups in BASH_REMATCH, quoting rules and pitfalls.
Bash can match a string against a regular expression without calling grep or sed. The operator =~ inside a [[ ... ]] test treats its right-hand side as a POSIX extended regular expression (ERE). If the string matches, the test succeeds, and Bash stores the whole match and every parenthesized group in the array BASH_REMATCH.
This is the cheapest way to validate input or pull a few fields out of one line, because no extra process is started. For anything line-by-line over a big file, grep or awk is still the better tool.
[[ STRING =~ REGEX ]] echo "${BASH_REMATCH[0]}" # whole match echo "${BASH_REMATCH[1]}" # first (...) group
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.
Syntax
[[ string =~ pattern ]]
| Name | Description |
|---|---|
string |
The text to search. It is not word-split, so quoting is optional. |
pattern |
An extended regular expression. The match is unanchored: it succeeds if the pattern matches anywhere. Add ^ and $ to anchor. |
BASH_REMATCH[0] |
The part of the string that matched the whole pattern. |
BASH_REMATCH[n] |
The text matched by the n-th parenthesized group; empty if that group did not take part. |
BASH_REMATCH after failure |
Empty (no elements). |
Returns: 0 for a match, 1 for no match, 2 if the regular expression is syntactically invalid.
Errors and edge cases: the regex flavor is whatever the C library's regcomp supports, so there are no lazy quantifiers (.*?), no lookahead and no \d. Use [0-9] and [[:digit:]] instead.
Usage notes
- Quoting rules changed once. In Bash 3.1 a quoted pattern was still treated as a regex; from 3.2 on, quoted parts are literal text and unquoted parts are regex. The portable habit is to store the pattern in a variable and use it unquoted:
[[ $x =~ $re ]]. - Spaces and special characters. An unquoted space ends the pattern, so a pattern with spaces must be quoted or put in a variable. Write
[[:space:]]instead of a literal space when you can. - Parentheses and
|. In a pattern given directly inside[[ ]], bare()|are accepted, unlike in the rest of the shell, but quoting them would make them literal. A variable avoids any doubt. - Capturing repeated groups. A group inside a quantifier, such as
(a|b)+, keeps only the last repetition inBASH_REMATCH[1]. - Position info.
BASH_REMATCHhas no offsets. To find where something matched, strip the prefix from the string yourself, as in the example above. - Not for finding every match. There is no global flag, so loop and consume the string, or use
grep -o. nocasematch(shopt -s nocasematch) makes=~,==in[[ ]]andcaseignore case; remember to turn it off again.- The array is global. The next
=~overwrites it, so copy values into named variables right away if you need them later.