CODEWITHSUNDEEP
c#asp.net.netstreamgzip
Matias Cicero
Matias Cicero

Reputation: 26281

Compressing with GZipStream results in more bytes

I am using the following methods to compress my response's content:

(Consider _compression = CompressionType.GZip)

private async Task<HttpContent> CompressAsync(HttpContent content)
{
     if(content == null) return null;
     byte[] compressedBytes;
     using(MemoryStream outputStream = new MemoryStream())
     {
         using (Stream compressionStream = GetCompressionStream(outputStream))
         using (Stream contentStream = await content.ReadAsStreamAsync())
             await contentStream.CopyToAsync(compressionStream);

         compressedBytes = outputStream.ToArray();
     }

     content.Dispose();

     HttpContent compressedContent = new ByteArrayContent(compressedBytes);
     compressedContent.Headers.ContentEncoding.Add(GetContentEncoding());
     return compressedContent;
}

private Stream GetCompressionStream(Stream output)
{
    switch (_compression)
    {
        case CompressionType.GZip: { return new GZipStream(output, CompressionMode.Compress); }
        case CompressionType.Deflate: { return new DeflateStream(output, CompressionMode.Compress); }
        default: return null;
    }
}

private string GetContentEncoding()
{
    switch (_compression)
    {
        case CompressionType.GZip: { return "gzip"; }
        case CompressionType.Deflate: { return "deflate"; }
        default: return null;
    }
}

However, this method returns more bytes than the original content.

For example, my initial content is 42 bytes long, and the resulting compressedBytes array has a size of 62 bytes.

Am I doing something wrong here? How can compression generate more bytes?

Upvotes: 1

Views: 741

Answers (3)

michalk
michalk

Reputation: 566

For such a small size the compression overhead can actually make the file larger and it's nothing unusual. Here it's explained with more detail.

Upvotes: 0

ngu
ngu

Reputation: 894

When compressing small files with gzip it is possible that the metadata (for the compressed file itself) causes an increase larger then the number of bytes saved by compression.

See Googles gzip tips:

Believe it or not, there are cases where GZIP can increase the size of the asset. Typically, this happens when the asset is very small and the overhead of the GZIP dictionary is higher than the compression savings, or if the resource is already well compressed.

Upvotes: 1

sstan
sstan

Reputation: 36483

You are not necessarily doing anything wrong. You have to take into account that these compressed formats always require a bit of space for header information. So that's probably why it grew by a few bytes.

Under normal circumstances, you would be compressing larger amounts of data. In that case, the overhead associated with the header data becomes unnoticeable when compared to the gains you make by compressing the data.

But because, in this case, your uncompressed data is so small, and so you are probably not gaining much in the compression, then this is one of the few instances where you can actually notice the header taking up space.

Upvotes: 1

Related Questions