Bash Regex - Pattern Matching with =~, grep, sed, and awk

Regular Expressions in Bash

Bash supports regular expressions through the =~ operator in [[ ]] tests, and through external commands like grep, sed, and awk.

The =~ Operator

The =~ operator tests a string against a regex pattern:

# Basic match
if [[ "hello123" =~ [0-9]+ ]]; then
  echo "Contains numbers"
fi

# Capture groups with BASH_REMATCH
VERSION="v3.14.2"
if [[ "$VERSION" =~ ^v([0-9]+)\.([0-9]+)\.([0-9]+)$ ]]; then
  echo "Major: ${BASH_REMATCH[1]}"  # 3
  echo "Minor: ${BASH_REMATCH[2]}"  # 14
  echo "Patch: ${BASH_REMATCH[3]}"  # 2
fi

# Negated match
if [[ ! "$input" =~ ^[0-9]+$ ]]; then
  echo "Not a number"
fi

Common Validation Patterns

# Email validation
[[ "$email" =~ ^[a-zA-Z0-9._%+-]+@[a-zA-Z0-9.-]+\.[a-zA-Z]{2,}$ ]]

# IPv4 address
[[ "$ip" =~ ^([0-9]{1,3}\.){3}[0-9]{1,3}$ ]]

# Semantic version
[[ "$ver" =~ ^[0-9]+\.[0-9]+\.[0-9]+$ ]]

# URL
[[ "$url" =~ ^https?:// ]]

# Date (YYYY-MM-DD)
[[ "$date" =~ ^[0-9]{4}-[0-9]{2}-[0-9]{2}$ ]]

grep with Regex

# Extended regex (-E)
grep -E "error|warning|fatal" app.log

# Match IP addresses
grep -oE "([0-9]{1,3}\.){3}[0-9]{1,3}" access.log

# Match email addresses
grep -oE "[a-zA-Z0-9._%+-]+@[a-zA-Z0-9.-]+\.[a-zA-Z]{2,}" contacts.txt

# Perl-compatible regex (-P)
grep -P "\d{3}-\d{2}-\d{4}" data.txt

sed with Regex

# Capture group back-references
echo "John Smith" | sed -E 's/^(\w+) (\w+)$/\2, \1/'
# Output: Smith, John

# Remove HTML tags
sed 's/<[^>]*>//g' page.html

# Extract between delimiters
echo "key=value" | sed 's/.*=//'

awk with Regex

# Match lines with a pattern
awk '/ERROR/ {print $0}' app.log

# Match a specific field
awk -F',' '$3 ~ /^[0-9]+$/ {print $1, $3}' data.csv

# Substitute in a field
awk '{gsub(/old/, "new", $2); print}' data.txt