Modifying a File’s Content Based on Patterns in Java

1. Overview

When maintaining or transforming text data, we often need to modify the content of existing files. In Java, there are multiple ways to accomplish this, depending on the file size and our performance needs. Small files can be easily handled in memory, while large files are better processed line by line using streams.

In this tutorial, we’ll explore two approaches to modify a file’s content based on patterns. Both approaches rely on Java’s modern Java NIO APIs.

2. Introduction to the Problem

Before jumping into code, let’s define the problem we want to solve.

2.1. Defining the Problem

Imagine that we are given a text file containing lines about various programming languages. Here’s what it looks like:

Both JAVA and KOTLIN applications can run on the JVM.
But python is a simpler language
PYTHON application is also platform independent.
java and kotlin are statically typed languages.
On the other hand, python is a dynamically typed language.

We want to clean up and normalize this content by performing several text transformations:

• Remove the second line.
• Normalize capitalization – Convert all occurrences of “java”, “kotlin”, and “python” to their correctly capitalized forms “Java“, “Kotlin“, and “Python“
• Expand abbreviations – Wherever “JVM” appears, add its complete form “(Java Virtual Machine)” after it.
• Write back the changes to the original file.

At the end, our modified file should look like this:

Both Java and Kotlin applications can run on the JVM (Java Virtual Machine).
Python application is also platform independent.
Java and Kotlin are statically typed languages.
On the other hand, Python is a dynamically typed language.

As usual, we’ll use unit test methods to demonstrate each approach. So next, let’s set up our unit test class.

2.2. Setting up the Test

Let’s first look at the unit test setup:

public class ModifyFileByPatternUnitTest {
    @TempDir
    private File fileDir;

    private File myFile;

    private static final List<String> ORIGINAL_LINES = List.of(
      "Both JAVA and KOTLIN applications can run on the JVM.",
      "But python is a simpler language",
      "PYTHON application is also platform independent.",
      "java and kotlin are statically typed languages.",
      "On the other hand, python is a dynamically typed language.");

    private static final List<String> EXPECTED_LINES = List.of(
      "Both Java and Kotlin applications can run on the JVM (Java Virtual Machine).",
      "Python application is also platform independent.",
      "Java and Kotlin are statically typed languages.",
      "On the other hand, Python is a dynamically typed language.");

    @BeforeEach
    void initFile() throws IOException {
        myFile = new File(fileDir, "myFile.txt");
        Files.write(myFile.toPath(), ORIGINAL_LINES);
    }

    // ...
}

As we can see, to ensure that our tests don’t modify any real files on the system, we’ll take advantage of the @TempDir annotation, which allows us to create a temporary directory that exists only for the duration of the test. Once the test finishes, JUnit automatically deletes all files inside it. This provides a clean, isolated environment for every run.

Additionally, the initFile() method ensures that every test begins with the same initial file content, thereby ensuring our modifications are tested in a controlled environment.

Next, let’s see how to solve this problem.

3. Loading the Entire File into Memory

Our first approach is to load the entire file into memory, modify its contents, and write it back. This technique is simple and effective for small or moderately sized files.

Let’s check how this is done:

@Test
void whenLoadTheFileContentToMemModifyThenWriteBack_thenCorrect() throws IOException {
    assertTrue(Files.exists(myFile.toPath()));
    List<String> lines = Files.readAllLines(myFile.toPath());
    lines.remove(1); // remove the 2nd line
    List<String> newLines = lines.stream()
      .map(line -> line.replaceAll("(?i)java", "Java")
        .replaceAll("(?i)kotlin", "Kotlin")
        .replaceAll("(?i)python", "Python")
        .replaceAll("JVM", "$0 (Java Virtual Machine)"))
      .toList();
    Files.write(myFile.toPath(), newLines);
    assertLinesMatch(EXPECTED_LINES, Files.readAllLines(myFile.toPath()));
}

In this approach, we first read all lines from the file into a List using Files.readAllLines(). Then, we removed the second line and performed the pattern replacements. Finally, we used Files.write() overwrites the file directly with the updated lines.

It’s worth mentioning that we used JUnit 5’s assertLinesMatch() to verify the file content.

This in-memory approach is concise and easy to reason about. However, it’s not ideal for huge files where memory usage becomes a concern.

So next, let’s assume myFile.txt is a huge file and explore how to solve the problem.

4. Memory-Efficient File Modification Using BufferedReader

To handle larger files efficiently, we can use streaming IO with BufferedReader and BufferedWriter. This method processes one line at a time without loading the entire file into memory.

Next, let’s have a look at the implementation:

@Test
void whenUsingBufferedReaderAndModifyViaTempFile_thenCorrect(@TempDir Path tempDir) throws IOException {
    Pattern javaPat = Pattern.compile("(?i)java");
    Pattern kotlinPat = Pattern.compile("(?i)kotlin");
    Pattern pythonPat = Pattern.compile("(?i)python");
    Pattern jvmPat = Pattern.compile("JVM");

    Path modifiedFile = tempDir.resolve("modified.txt");

    try (BufferedReader reader = Files.newBufferedReader(myFile.toPath());
      BufferedWriter writer = Files.newBufferedWriter(modifiedFile)) {

        int lineNumber = 0;
        String line;
        while ((line = reader.readLine()) != null) {
            lineNumber++;
            if (lineNumber == 2) {
                continue; // skip the 2nd line
            }

            String replaced = line;
            replaced = javaPat.matcher(replaced)
              .replaceAll("Java");
            replaced = kotlinPat.matcher(replaced)
              .replaceAll("Kotlin");
            replaced = pythonPat.matcher(replaced)
              .replaceAll("Python");
            replaced = jvmPat.matcher(replaced)
              .replaceAll("JVM (Java Virtual Machine)");

            writer.write(replaced);
            writer.newLine();
        }
    }

    Files.move(modifiedFile, myFile.toPath(), StandardCopyOption.REPLACE_EXISTING);
    assertTrue(myFile.exists());
    assertLinesMatch(EXPECTED_LINES, Files.readAllLines(myFile.toPath()));
}

Next, let’s understand the details.

• Precompiled Regex Patterns – Instead of calling replaceAll() directly on Strings, we compile the Regex patterns once using Pattern.compile(). This is more efficient when applying the same patterns multiple times.
• Streaming with BufferedReader and BufferedWriter – We read the file line by line, modify each one, and write it immediately to a temporary output file (modified.txt). This keeps memory usage low regardless of file size.
• Try-with-resources – The try (…) { … } syntax automatically closes both reader and writer when the block finishes. It’s a safer and cleaner alternative to manually closing streams in a finally block.
• Replacing the original file – Once writing is complete, we replace the original file with the modified one using Files.move() and the REPLACE_EXISTING option.

This method is robust, scalable, and well-suited for large files or streaming scenarios that require fine-grained control over processing.

5. Conclusion

In this article, we explored two practical techniques for modifying a file’s content in Java based on regular expression patterns:

• Loading the entire file into memory — It’s a simple and expressive solution, ideal for smaller files.
• Streaming with BufferedReader and BufferedWriter — It’s an efficient and scalable approach for handling large files.

Both methods build upon Java’s standard libraries, so they work in any Java environment without extra dependencies. By combining these approaches, we can handle a wide range of file transformation tasks safely and efficiently.

As always, the complete source code for the examples is available over on GitHub.