How do you find and replace text in multiple files using shell scripting?

Using shell scripts to search and replace text across multiple files is a common task, typically achieved with command-line tools such as sed (stream editor). The sed utility is a powerful text processing tool designed to process data from standard input or a series of files and output the results.

Here is a simple example demonstrating how to use the sed command to search and replace text across multiple files:

Suppose we want to search for the word 'error' and replace it with 'warning' in all .txt files within a project. We can use the following shell script command:

bash
#!/bin/bash

# Define the search and replacement text
search="error"
replace="warning"

# Navigate to the directory containing the target files
cd /path/to/files

# Perform search and replace on all .txt files in the directory
for file in *.txt
do
  sed -i "s/$search/$replace/g" "$file"
done

echo "Replacement completed!"

In this script:

• The -i option of the sed command directly modifies the file content.

• The s/$search/$replace/g is a sed replacement command where s denotes substitution with the format s/pattern/replacement/flags:

  • pattern is the text to search for, represented by the variable $search.
  • replacement is the new text to replace, represented by the variable $replace.
  • flags include g for global replacement, meaning all occurrences on a line are replaced.

• The for file in *.txt loop iterates through all .txt files in the directory.

Additionally, caution is advised when using this method, as incorrect specification of search and replacement text may cause unnecessary data loss. It is recommended to test the script on a small number of files before executing the replacement to ensure it functions as expected.