Find

Enhance your text-processing workflow by using sed to search, print, or delete specific patterns directly within files. While similar to grep, sed lets you combine pattern matching with editing commands in one step.

To follow along, we’ll use an employees.txt file with records formatted as ID|First|Last|Department|Role|Email|Salary.

1. Basic Search Syntax

sed requires both a pattern and an action. The minimal form is:

sed '/pattern/' file

Without an explicit action, sed may default to printing every line or throw an error:

$ sed '/Manager/' employees.txt
# (No command given; behavior depends on your sed version)

To print only matching lines, use -n (quiet mode) with the p command:

sed -n '/Manager/p' employees.txt

• -n : suppress automatic printing
• /Manager/ : search for “Manager”
• p : print matching lines

$ sed -n '/Manager/p' employees.txt
5|Feng|Lin|Sales|Sales Manager|[email protected]|90000
6|Andy|Luscomb|IT|IT Manager|[email protected]|95000
7|Mark|Crocker|HR|HR Manager|[email protected]|85000
8|Jing|Ma|Engineering|Engineering Manager|[email protected]|100000

$ sed -n '/manager/p' employees.txt
# No match

2. Searching Substrings

Match partial strings by specifying only a fragment:

$ sed -n '/Ma/p' employees.txt
4|Enrique|Rivera|Marketing|Marketing Specialist|[email protected]|65000
5|Feng|Lin|Sales|Sales Manager|[email protected]|90000
6|Andy|Luscomb|IT|IT Manager|[email protected]|95000
7|Mark|Crocker|HR|HR Manager|[email protected]|85000
8|Jing|Ma|Engineering|Engineering Manager|[email protected]|100000

This matches “Marketing”, “Manager”, “Mark”, and the last name “Ma”.

3. Exact Word Matches with Word Boundaries

Use \< and \> to match whole words only:

sed -n '/\<Ma\>/p' employees.txt

$ sed -n '/\<Ma\>/p' employees.txt
8|Jing|Ma|Engineering|Engineering Manager|[email protected]|100000

Command breakdown:

• sed : invoke stream editor
• -n : suppress default output
• /\<Ma\>/ : match exact word “Ma”
• p : print matching line

4. Combining Multiple Search Patterns

Chain multiple scripts with -e to search for more than one pattern:

sed -n \
    -e '/\<Manager\>/p' \
    -e '/\<IT\>/p' \
    employees.txt

$ sed -n -e '/\<Manager\>/p' -e '/\<IT\>/p' employees.txt
3|Debbie|Miller|IT|Software Developer|[email protected]|80000
5|Feng|Lin|Sales|Sales Manager|[email protected]|90000
6|Andy|Luscomb|IT|IT Manager|[email protected]|95000
7|Mark|Crocker|HR|HR Manager|[email protected]|85000
8|Jing|Ma|Engineering|Engineering Manager|[email protected]|100000

Lines matching either pattern appear once per script.

5. Deleting Matches

Swap p for d to remove matching lines:

Remove Enrique’s record:

sed -e '/\<Enrique\>/d' employees.txt

Delete all lines containing “Ma” as a whole word:

$ sed -e '/\<Ma\>/d' employees.txt
1|Kriti|Shreshtha|Finance|Financial Analyst|[email protected]|60000
2|Rajasekar|Vasudevan|Finance|Senior Accountant|[email protected]|75000
3|Debbie|Miller|IT|Software Developer|[email protected]|80000
4|Enrique|Rivera|Marketing|Marketing Specialist|[email protected]|65000
5|Feng|Lin|Sales|Sales Manager|[email protected]|90000
6|Andy|Luscomb|IT|IT Manager|[email protected]|95000
7|Mark|Crocker|HR|HR Manager|[email protected]|85000

6. Editing Files In-Place with -i

Apply deletions or substitutions directly using -i:

$ sed -i \
    -e '/\<Manager\>/d' \
    -e '/\<IT\>/d' \
    employees.txt

$ cat employees.txt
1|Kriti|Shreshtha|Finance|Financial Analyst|[email protected]|60000
2|Rajasekar|Vasudevan|Finance|Senior Accountant|[email protected]|75000
4|Enrique|Rivera|Marketing|Marketing Specialist|[email protected]|65000

7. Quick Reference: sed Flags

Links and References

• GNU sed Manual
• Regular Expressions Basics
• grep vs sed Comparison