How to Remove Special Characters

Quick Answer: Remove Special Characters from a String

To remove special characters from a string in Bash, use the tr command: echo "$text" | tr -cd '[:alnum:]'. This keeps only letters and numbers, stripping everything else. For more control over what to keep, use sed or parameter expansion with regex.

Quick Comparison: Special Character Removal Methods

Method	Speed	Best For	Complexity
tr -cd	Very fast	Keep specific character sets	Simple
sed regex	Fast	Pattern-based removal	Moderate
Parameter expansion	Very fast	Simple patterns	Moderate
grep -o	Fast	Extracting matches	Simple
awk gsub	Fast	Complex transformations	Moderate

Bottom line: Use tr for simple character set filtering, use sed for pattern-based removal.

---

Removing Special Characters from Strings

Sometimes you need to clean strings by removing special characters, keeping only letters, numbers, or specific patterns. This is useful for sanitizing user input, creating safe filenames, or processing data.

Using tr Command

The simplest approach with tr:

#!/bin/bash

text="Hello-World_123!"

# Keep only alphanumeric
clean=$(echo "$text" | tr -cd '[:alnum:]')
echo "$clean"  # Output: HelloWorld123

Common tr Patterns

#!/bin/bash

text="Price: $99.99!"

# Remove punctuation
clean=$(echo "$text" | tr -d '[:punct:]')
echo "$clean"  # Output: Price 9999

# Remove all non-alphanumeric
clean=$(echo "$text" | tr -cd '[:alnum:] ')
echo "$clean"  # Output: Price 9999

# Remove all non-digits
digits=$(echo "$text" | tr -cd '[:digit:]')
echo "$digits"  # Output: 9999

Using sed

More powerful pattern removal:

#!/bin/bash

text="hello@world#bash!"

# Remove all special chars except spaces
clean=$(echo "$text" | sed 's/[^a-zA-Z0-9 ]//g')
echo "$clean"  # Output: helloworld bash

Practical Example: Sanitize Filename

#!/bin/bash

sanitize_filename() {
  local filename="$1"
  # Keep only alphanumeric, dash, underscore, dot
  local clean=$(echo "$filename" | tr -cd '[:alnum:]._-')
  echo "$clean"
}

result=$(sanitize_filename "My Document (2026).pdf")
echo "Safe filename: $result"  # Output: MyDocument2026.pdf

Real-World Example: Clean User Input

#!/bin/bash

# Remove special chars from username input
clean_username() {
  local input="$1"
  # Keep only lowercase letters and numbers
  echo "$input" | tr '[:upper:]' '[:lower:]' | tr -cd '[:alnum:]_'
}

username=$(clean_username "John_Doe@123")
echo "Username: $username"  # Output: john_doe123

Bash Parameter Expansion Method

Pure Bash without external commands:

#!/bin/bash

remove_special() {
  local text="$1"
  local clean=""

  for ((i=0; i<${#text}; i++)); do
    char="${text:$i:1}"
    # Keep only alphanumeric
    if [[ "$char" =~ [[:alnum:]] ]]; then
      clean+="$char"
    fi
  done

  echo "$clean"
}

result=$(remove_special "Hello@World!")
echo "$result"  # Output: HelloWorld

Comparison of Methods

text="test@123!"

# tr method (fast)
echo "$text" | tr -cd '[:alnum:]'

# sed method (more flexible)
echo "$text" | sed 's/[^a-zA-Z0-9]//g'

# Bash loop (no external tools, slower)

Important Notes

• tr -d deletes characters
• tr -cd keeps only specified characters
• [:alnum:] includes letters and digits
• [:punct:] includes punctuation marks
• Use [^...] in regex to negate patterns

Summary

Use tr for fast character removal, sed for complex patterns, or pure Bash loops when no external commands are available. The tr -cd pattern is most efficient for keeping only safe characters.