How to Extract Numbers from String

Quick Answer: Extract Numbers from String in Bash

To extract numbers from a string, use grep -oE '[0-9]+' for integers or grep -oE '[0-9]+\.[0-9]+' for decimals. For a single variable without pipes, use regex: [[ "$text" =~ ([0-9]+) ]] then access ${BASH_REMATCH[1]}.

Quick Comparison: Number Extraction Methods

Method	Syntax	Best For	Speed
grep -oE	grep -oE '[0-9]+'	Pipes/files	Fast
Regex match	[[ =~ ([0-9]+) ]]	Variables	Fastest
sed -n s	sed -n 's/.*\([0-9]*\).*/\1/p'	Piped data	Medium
awk match	awk '{match($0, /[0-9]+/); print substr($0, RSTART, RLENGTH)}'	Fields	Medium

Bottom line: Use grep -oE for pipes; use regex capture for variables.

---

Extract numeric values from text strings using grep, sed, awk, and regex patterns.

Method 1: Using grep -o (Simplest)

# Extract all numbers
echo "Price: $19.99" | grep -o "[0-9]*"

# Extract floating point
echo "Price: $19.99" | grep -o "[0-9.]*"

# Extract with leading zeros
echo "ID: 00123" | grep -o "[0-9]*"

Using sed

# Remove non-digits
echo "abc123def456" | sed 's/[^0-9]//g'
# Output: 123456

# Extract first number
echo "Price: $19.99" | sed 's/[^0-9.]*//g'

Using awk

# Extract fields with numbers
echo "Item: 5 Price: 19.99" | awk '{print $2, $4}'

# Extract only numeric fields
echo "abc 123 def 456" | awk '{for(i=1;i<=NF;i++) if($i~/^[0-9]+$/) print $i}'

Practical Example: Extract All Numbers

#!/bin/bash

text="$1"

# Extract all integers
numbers=$(echo "$text" | grep -oE '[0-9]+')

echo "Found numbers:"
echo "$numbers"

Usage:

$ ./extract_numbers.sh "The prices are 10, 20, 30 dollars"
Found numbers:
10
20
30

Extract Numbers from Variables

#!/bin/bash

string="Order #12345 total: $99.99"

# Remove non-digits
order=$(echo "$string" | sed 's/[^0-9]//g')
echo "All digits: $order"

# Extract first number only
first=$(echo "$string" | grep -oE '[0-9]+' | head -1)
echo "First number: $first"

# Extract decimal
price=$(echo "$string" | grep -oE '[0-9]+\.[0-9]+')
echo "Price: $price"

Output:

All digits: 12345999.99
First number: 12345
Price: 99.99

Using Bash Parameter Expansion

#!/bin/bash

string="The answer is 42!"

# Remove leading non-digits
result="${string##*([^0-9])}"
result="${result%%[^0-9]*}"

echo "$result"  # Output: 42

Extract with Regex Capture

#!/bin/bash

text="Invoice #INV-2026-001 for $500.00"

# Extract invoice number using regex
if [[ $text =~ INV-([0-9]+)-([0-9]+) ]]; then
  year="${BASH_REMATCH[1]}"
  num="${BASH_REMATCH[2]}"
  echo "Invoice: $year-$num"
fi

# Extract price
if [[ $text =~ \$([0-9]+\.[0-9]+) ]]; then
  price="${BASH_REMATCH[1]}"
  echo "Amount: $price"
fi

Output:

Invoice: 2026-001
Amount: 500.00

Extract and Sum

#!/bin/bash

# Extract all numbers and sum them
text="Scores: 95, 87, 92, 88"

sum=$(echo "$text" | grep -oE '[0-9]+' | awk '{sum+=$1} END {print sum}')
echo "Total: $sum"

Output:

Total: 362

Multiple Patterns

#!/bin/bash

# Extract multiple types of numbers
text="Phone: 555-123-4567 Date: 2026-02-21 Price: $99.99"

# Phone
phone=$(echo "$text" | grep -oE '[0-9]{3}-[0-9]{3}-[0-9]{4}')

# Date
date=$(echo "$text" | grep -oE '[0-9]{4}-[0-9]{2}-[0-9]{2}')

# Price
price=$(echo "$text" | grep -oE '\$[0-9]+\.[0-9]{2}')

echo "Phone: $phone"
echo "Date: $date"
echo "Price: $price"

From File Lines

#!/bin/bash

# Extract numbers from each line
file="data.txt"

while IFS= read -r line; do
  # Get first number from line
  num=$(echo "$line" | grep -oE '[0-9]+' | head -1)
  echo "Line: $line → Number: $num"
done < "$file"

Common Mistakes

1. Using wrong regex - [0-9]* matches zero digits
2. Not escaping dots - use \. for literal dot
3. Not quoting strings - failures with spaces
4. Wrong grep flags - use -o to get match only

Key Points

• Use [0-9]+ for one or more digits
• Use [0-9.]+ for decimals
• Use -o flag with grep
• Quote variables properly
• Test patterns with sample data

Summary

Extracting numbers from strings is common in shell scripts. Use grep -o for simple cases, awk for field-based extraction, and regex with capture groups for complex patterns.