Basic Regular Expressions in Bash ([[ =~ ]])

Basic Pattern Matching

The =~ operator inside [[ ]] tests whether the left-hand string matches the extended regular expression on the right. The exit status of the [[ ]] reflects whether the match succeeded.

Example: Basic Pattern Matching

#!/bin/bash
value="hello123"
if [[ $value =~ [0-9]+ ]]; then
    echo "contains digits"
fi

Capturing Groups with BASH_REMATCH

When a regex includes parenthesized groups, a successful =~ match populates the BASH_REMATCH array: index 0 is the whole match, and indices 1+ correspond to each capture group in order.

Example: Capturing Groups with BASH_REMATCH

#!/bin/bash
version="v2.5.1"
if [[ $version =~ v([0-9]+)\.([0-9]+)\.([0-9]+) ]]; then
    echo "major: ${BASH_REMATCH[1]}"
    echo "minor: ${BASH_REMATCH[2]}"
    echo "patch: ${BASH_REMATCH[3]}"
fi

Storing a Pattern in a Variable

Keeping the regex in a variable and using it unquoted in the [[ ]] test avoids accidental literal-string matching and makes complex patterns easier to read and reuse.

Warning: Quoting the pattern (e.g. [[ $email =~ "$pattern" ]]) turns it into a literal string match in bash, silently breaking the regex.

Example: Storing a Pattern in a Variable

#!/bin/bash
email="[email protected]"
pattern='^[A-Za-z0-9._%+-]+@[A-Za-z0-9.-]+\.[A-Za-z]{2,}$'
if [[ $email =~ $pattern ]]; then
    echo "looks like a valid email format"
else
    echo "does not look like an email"
fi

Validating Input Format

Regex matching with =~ is commonly used to validate that input has an expected shape, such as being all digits, before proceeding with further processing.

Example: Validating Input Format

#!/bin/bash
check_number() {
    if [[ $1 =~ ^[0-9]+$ ]]; then
        echo "$1 is a valid whole number"
    else
        echo "$1 is NOT a valid whole number"
    fi
}
check_number "12345"
check_number "12a45"