Substitute

The s (substitute) command in sed is a powerful tool for search-and-replace operations on text streams. Whether you’re updating configuration files, automating code refactoring, or cleaning log files, mastering sed substitution will streamline your workflow.

Table of Contents

1. Basic Syntax
2. Quick Reference Commands
3. Practical Example: Updating a Salary
4. Substitution Scope
5. Targeting Specific Occurrences
6. Line Addressing
7. In-Place Editing (-i)
8. Inserting Text with i
9. Conclusion
10. References

Basic Syntax

Use the following pattern to substitute old_string with new_string:

sed 's/old_string/new_string/' input-file

• s     — substitute command
• old_string — search pattern (regular expression supported)
• new_string — replacement text
• /     — delimiters (can be any character)
• By default, only the first match per line is replaced.

Quick Reference Commands

Practical Example: Updating a Salary

Given an employees.txt file:

To update Enrique Rivera’s salary from 65000 to 85000, run:

sed 's/65000/85000/' employees.txt

This command replaces the first occurrence of 65000 on each line and prints the result to stdout.

Substitution Scope

Global Replacement

Append the g flag to replace all matches in each line:

sed 's/\bIT\b/Information Technology/g' employees.txt

First-Match Only

Without g, only the first match is replaced:

sed 's/\bIT\b/Information Technology/' employees.txt

Targeting Specific Occurrences

You can replace only the nth occurrence on each line by specifying a number:

sed 's/\bIT\b/Information Technology/2' employees.txt

The above replaces only the second IT per line.

Line Addressing

Limit substitutions to certain lines or ranges:

• Single line (line 7):

  sed '7 s/\bHR\b/Human Resources/' employees.txt
  

• Line range (lines 1–3):

  sed '1,3 s/Finance/Taxes/' employees.txt
  

• Another single line (line 5):

  sed '5 s/\bSales\b/Transactions/' employees.txt
  

In-Place Editing (-i)

Modify the file directly using -i:

sed -i 's/\bcompany\b/KodeKloud/g' employees.txt

This updates all instances of company in employees.txt.

sed -i.bak 's/old/new/g' file.txt

Inserting Text with i

The i command inserts lines before the current pattern space or at a specified line:

sed '1iID|Name|FirstName|LastName|Job|Department|Email|Salary' employees.txt

Output:

ID|Name|FirstName|LastName|Job|Department|Email|Salary
1|Kriti|Shreshtha|Finance|Financial Analyst|[email protected]|60000
2|Rajasekar|Vasudevan|Finance|Senior Accountant|[email protected]|75000
...

If you omit the text after i, sed throws an error:

$ sed 'i' employees.txt
sed: -e expression #1, char 1: expected \ after 'a', 'c' or 'i'

Conclusion

In this guide, you learned how to:

• Use s/old/new/ for basic substitution
• Leverage g and numeric flags for global or specific replacements
• Address lines and ranges for targeted edits
• Apply in-place editing with -i (and backups)
• Insert new lines using the i command

Together, these techniques form the foundation of efficient text processing with sed.

References

• GNU sed Manual
• Linux sed Cheat Sheet
• Regular Expressions in sed