How to Use Sed to Replace | ActiveDirectoryTools

Quick Answer: Replace Text with Sed

To replace text, use sed 's/old/new/' file.txt for the first occurrence per line or sed 's/old/new/g' file.txt for all occurrences. For in-place editing, use sed -i 's/old/new/g' file.txt. The ‘s’ command is sed’s substitution command.

Quick Comparison: Sed Replacement Methods

Syntax	What It Does	Best For	Notes
s/old/new/	First on line	Basic replace	Simplest
s/old/new/g	All on line	Global replace	Most common
s/old/new/gi	All, case-insensitive	Ignoring case	Useful for mixed case
-i flag	In-place edit	Modifying files	Creates backup with -i.bak

Bottom line: Use s/old/new/g for global replacement; add i for case-insensitive.

Use sed (stream editor) to find and replace text in files and streams. Learn basic substitution, advanced patterns, and file editing.

Method 1: Basic Substitution

# Replace first occurrence per line
sed 's/old/new/' file.txt

# Replace all occurrences per line
sed 's/old/new/g' file.txt

# Show on stdout (doesn't modify file)

Detailed Example

Test file (data.txt):

apple is red
banana is yellow
apple is sweet

# Replace first "apple" on each line
sed 's/apple/orange/' data.txt

# Output:
# orange is red
# banana is yellow
# orange is sweet

# Replace all occurrences
sed 's/apple/orange/g' data.txt

# Output:
# orange is red
# banana is yellow
# orange is sweet

In-Place File Editing

# Modify file directly (with backup)
sed -i.bak 's/old/new/g' file.txt

# Modify without backup
sed -i 's/old/new/g' file.txt

Case-Insensitive Replacement

# Case-insensitive flag (i)
sed 's/apple/orange/gi' file.txt

Replace on Specific Lines

# Replace only on line 2
sed '2s/old/new/' file.txt

# Replace on lines 2-4
sed '2,4s/old/new/' file.txt

# Replace on line 5 onwards
sed '5,$s/old/new/' file.txt

Replace with Pattern Matching

# Only replace on lines containing "pattern"
sed '/pattern/s/old/new/' file.txt

# Replace on lines matching regex
sed '/^Error:/s/failed/error/' logfile.txt

# Replace if line contains specific text
sed '/^[0-9]/s/^/#/' file.txt  # Comment out lines starting with digit

Use Different Delimiter

# Use | instead of / (useful for paths)
sed 's|/old/path|/new/path|' file.txt

# Use # as delimiter
sed 's#old#new#' file.txt

Practical Example: Configuration Update

#!/bin/bash

# File: update_config.sh

config_file="$1"

# Backup original
cp "$config_file" "$config_file.backup"

# Update multiple settings
sed -i 's/^debug=false/debug=true/' "$config_file"
sed -i 's/^port=8080/port=9000/' "$config_file"
sed -i 's/^host=localhost/host=0.0.0.0/' "$config_file"

echo "✓ Configuration updated"
echo "Backup saved: $config_file.backup"

Usage:

$ chmod +x update_config.sh
$ ./update_config.sh app.conf

Escape Special Characters

# For replacements with special regex characters, escape them
sed 's/\./\\./g' file.txt         # Escape dots
sed 's/\$/USD/g' file.txt         # Escape dollar sign
sed 's/\*/asterisk/g' file.txt    # Escape asterisk

Address Ranges with Substitution

# Replace between two patterns
sed '/START/,/END/s/old/new/' file.txt

# Replace on odd-numbered lines
sed -n '1~2p' file.txt             # Show odd lines
sed '1~2s/old/new/' file.txt       # Replace on odd lines

# Replace on even-numbered lines
sed '0~2s/old/new/' file.txt

Multiple Substitutions

# Using -e for multiple patterns
sed -e 's/apple/orange/' -e 's/banana/grape/' file.txt

# Or in script file
cat > replacements.sed << 'EOF'
s/apple/orange/
s/banana/grape/
s/cherry/lemon/
EOF

sed -f replacements.sed file.txt

Advanced Substitution with Captures

# Capture groups and backreferences
sed 's/\([a-z]*\) \([a-z]*\)/\2 \1/' file.txt

# Swap words: "hello world" becomes "world hello"
echo "hello world" | sed 's/\([^ ]*\) \([^ ]*\)/\2 \1/'

Print Modified Lines

# Show only lines that were changed
sed -n 's/old/new/p' file.txt

# Show all lines except those matching pattern
sed '/pattern/d' file.txt

Practical Example: Log File Processing

#!/bin/bash

# File: process_log.sh

logfile="$1"
output="${2:-processed.log}"

# Clean up log: remove timestamps, IPs, and sensitive data
sed \
  -e 's/\[.*\]/[LOG]/g' \
  -e 's/[0-9]\{1,3\}\.[0-9]\{1,3\}\.[0-9]\{1,3\}\.[0-9]\{1,3\}/IP_REDACTED/g' \
  -e 's/user[0-9]*/USER_X/g' \
  "$logfile" > "$output"

echo "Processed log saved to: $output"

Global vs Line-Specific

# Replace only first occurrence on line
sed 's/pattern/replacement/' file.txt

# Replace all occurrences on line
sed 's/pattern/replacement/g' file.txt

# Replace only second occurrence on line
sed 's/pattern/replacement/2' file.txt

# Replace from 2nd occurrence onwards
sed 's/pattern/replacement/2g' file.txt

Conditional Replacement

# Replace only if line contains another pattern
sed '/contains_this/ s/old/new/' file.txt

# Replace only if line does NOT contain pattern
sed '/skip_this/! s/old/new/' file.txt

Backup with Automatic Naming

# Create backup with timestamp
sed -i.$(date +%Y%m%d_%H%M%S) 's/old/new/g' file.txt

# Create backup with number
sed -i~1 's/old/new/g' file.txt  # Creates file.txt~1

Using Variables in sed

#!/bin/bash

old_value="$1"
new_value="$2"
file="$3"

# Must use double quotes to expand variables
sed -i "s/$old_value/$new_value/g" "$file"