---
title: "Compress and decompress bzip2 in C#"
url: https://recursiveai.org/onearchive/cookbook/compress-decompress-bzip2/
description: "Compress and decompress bzip2 in C# with a free, open-source .NET library. Run the complete example, restore the result and compare its contents."
version: "1.2.0"
---

# Compress and decompress bzip2 in C#

**OneArchive** is a free, open-source archive and compression library for .NET from Recursive AI: archives (zip, tar, 7z, RAR 5, cpio, ar, ISO) in RecursiveAI.OneArchive, codecs (zlib, gzip, xz, Zstandard, Brotli) in RecursiveAI.OneCompress. MIT or Apache-2.0, no native dependencies.

[NuGet](https://www.nuget.org/packages/RecursiveAI.OneCompress) · [GitHub](https://github.com/recursive-ai-oss/onearchive-csharp) · [Docs](https://recursiveai.org/onearchive/docs/) · [Download example (.zip)](https://recursiveai.org/onearchive/cookbook/compress-decompress-bzip2/compress-decompress-bzip2.zip)

Bzip2 compresses one byte stream into a `.bz2` file. Use it for a single file, or compress a tar archive when you need to keep several files together.

```shell
dotnet add package RecursiveAI.OneCompress
```

Requires [.NET 10](https://dotnet.microsoft.com/download/dotnet/10.0) (net10.0); .NET 9 and earlier aren’t supported.

```csharp
using RecursiveAI.OneCompress.BZip2;

byte[] packed = BZip2.Compress("Hello!"u8);
byte[] restored = BZip2.Decompress(packed);
Console.WriteLine(
    System.Text.Encoding.UTF8.GetString(restored));
```

```text
Hello!
```

Unzip [the example](https://recursiveai.org/onearchive/cookbook/compress-decompress-bzip2/compress-decompress-bzip2.zip), then `dotnet run`.

## The full program

The program generates `message.txt`: the text `Hello from BZip2!` followed by a newline, repeated 100 times. It writes `message.txt.bz2` and `restored.txt`, then compares the bytes. It replaces these sample files on each run.

```csharp
using RecursiveAI.OneCompress.BZip2;
using System.Text;

byte[] original = Encoding.UTF8.GetBytes(
    string.Concat(Enumerable.Repeat("Hello from BZip2!\n", 100)));
File.WriteAllBytes("message.txt", original);
byte[] packed = BZip2.Compress(original);
File.WriteAllBytes("message.txt.bz2", packed);
byte[] restored = BZip2.Decompress(
    File.ReadAllBytes("message.txt.bz2"),
    new BZip2DecompressionOptions { MaxOutputSize = 1024 * 1024 });
File.WriteAllBytes("restored.txt", restored);
Console.WriteLine($"Input: {original.Length} bytes");
Console.WriteLine($"Compressed: {packed.Length} bytes");
Console.WriteLine($"Restored: {restored.Length} bytes");
Console.WriteLine($"Bytes match: {original.SequenceEqual(restored)}");
```

`dotnet run` prints:

```text
Input: 1800 bytes
Compressed: 87 bytes
Restored: 1800 bytes
Bytes match: True
```

Invalid or truncated compressed data raises `CompressionFormatException`; exceeding a decoder limit raises `CompressionLimitException`.

## Variations

Choose block-size level 1 and cap restored output at 1 KiB.

```csharp
using RecursiveAI.OneCompress.BZip2;

byte[] packed = BZip2.Compress("Hello!"u8,
    new BZip2Options { Level = 1 });
byte[] restored = BZip2.Decompress(packed,
    new BZip2DecompressionOptions { MaxOutputSize = 1024 });
Console.WriteLine(restored.Length);
```

```text
6
```

## Related recipes

- [Compress and decompress lzip in C#](https://recursiveai.org/onearchive/cookbook/compress-decompress-lzip/)
- [Compress Brotli data in C#](https://recursiveai.org/onearchive/cookbook/compress-brotli-data/)
- [Create an ISO 9660 image in C#](https://recursiveai.org/onearchive/cookbook/create-iso9660-image/)

## The short answer

To compress and decompress bzip2 in C#, install `RecursiveAI.OneCompress` and use `BZip2.Compress` and `BZip2.Decompress` from `RecursiveAI.OneCompress.BZip2`. The first method returns a complete bzip2 stream. The second restores the uncompressed bytes. Write the compressed result with `File.WriteAllBytes` to make a `.bz2` file. The package is managed C# and needs no native codec installation.

The full program above deliberately reads back the file it wrote, rather than decoding the byte array still in memory. It compares the restored bytes with the original, so a successful run means the disk round trip preserved the data. The compression size printed here describes only this small, repetitive sample, not a predicted ratio for your files.

## How do I decompress a bzip2 file?

Read the file as bytes, pass them to `BZip2.Decompress`, then write or interpret the returned bytes. The full program does exactly that for `message.txt.bz2`. For text, use the same encoding that produced the original bytes. For images or other binary data, save the bytes directly; a text decoder is not part of decompression.

Set `BZip2DecompressionOptions.MaxOutputSize` to the largest uncompressed payload your application expects. The default is 1 GiB, not the size of the compressed file. A small compressed input can produce much more output. The one-call method allocates the returned array, so choose a cap that fits your application’s memory budget as well as its file-size requirements.

The [decoder](https://recursiveai.org/onearchive/docs/bzip2-decoder/) accepts concatenated bzip2 streams and joins their output. This matters for files produced by tools that concatenate independently compressed chunks: do not stop after the first chunk. Checksums are verified by default. Truncation, corruption and unexpected trailing bytes are errors, rather than a reason to quietly return partial data.

## How do I decompress a tar.bz2 file?

There are two layers: bzip2 compression on the outside, and tar entries on the inside. Calling `BZip2.Decompress` removes only the outer layer. Its output is still a tar archive, not a directory of extracted files. A `.bz2` extension by itself does not mean that the content is tar.

Use the archive API in `RecursiveAI.OneArchive` when you need to extract the files inside a compressed tar. It handles archive entries rather than just restoring a byte stream. The [compressed-tar extraction recipe](https://recursiveai.org/onearchive/cookbook/extract-tar-gz/) shows that API. Keep the distinction when naming your own output: compressing `message.txt` gives `message.txt.bz2`; it does not make a tar archive.

For writing, [bzip2 options](https://recursiveai.org/onearchive/docs/bzip2-encoder/) expose `Level` from 1 through 9, with 9 as the default. It controls the block-size level in the stream header. It is not a target file size. `MaxCodeLength` controls the longest Huffman code, with a default of 17 and an accepted range of 9 through 20. Start with defaults unless your application has a specific reason to change them; neither setting guarantees that arbitrary input becomes smaller.
