How to unzip a file in memory to an Azure CloudBlob

Question

Azure doesn't allow me to unzip a file directly in a container. I downloaded a zip file and now need to expand the file in the zip file. What I get is a 0 byte file. I can download the zip to my local computer and see the embedded csv, so the zip file isn't corrupt. I get no errors, just a zero byte output file. What am I doing wrong? I've tried all of these options unsuccessfully:

using (MemoryStream ms = new MemoryStream())
{
    await zipOutputBlob.DownloadToStreamAsync(ms);
    using (var zipStream = new GZipStream(ms, CompressionMode.Decompress))
    {
        CloudBlockBlob unzippedBlob = container.GetBlockBlobReference(String.Format("{0:yyyy-MM-dd}", lastWriteTime) + " " + Path.GetFileNameWithoutExtension(fileName) +  ".csv");
        unzippedBlob.Properties.ContentType = "text/csv";
        using (Stream outputFileStream = await unzippedBlob.OpenWriteAsync())
        {
            await zipStream.CopyToAsync(outputFileStream);
            outputFileStream.Flush();
        }
    }
}

2nd try:

using (MemoryStream ms = new MemoryStream())
{
    await zipOutputBlob.DownloadToStreamAsync(ms);
    using (var zipStream = new GZipStream(ms, CompressionMode.Decompress))
    {
        CloudBlockBlob unzippedBlob = container.GetBlockBlobReference(String.Format("{0:yyyy-MM-dd}", lastWriteTime) + " " + Path.GetFileNameWithoutExtension(fileName) + ".csv");
        unzippedBlob.Properties.ContentType = "text/csv";
        await unzippedBlob.UploadFromStreamAsync(zipStream);
    }
}

3rd

using (MemoryStream ms = new MemoryStream())
{
    await zipOutputBlob.DownloadToStreamAsync(ms);
    using (var zipStream = new GZipStream(ms, CompressionMode.Decompress))
    {
        CloudBlockBlob unzippedBlob = container.GetBlockBlobReference(String.Format("{0:yyyy-MM-dd}", lastWriteTime) + " " + Path.GetFileNameWithoutExtension(fileName) + ".csv");
        unzippedBlob.Properties.ContentType = "text/csv";
        using (Stream outputFileStream = await unzippedBlob.OpenWriteAsync())
        {
            await zipStream.CopyToAsync(outputFileStream);
            outputFileStream.Flush();
        }
    }
}

4th

using (MemoryStream ms = new MemoryStream())
{
    await zipOutputBlob.DownloadToStreamAsync(ms);
    using (DeflateStream decompressionStream = new DeflateStream(ms, CompressionMode.Decompress))
    {
        CloudBlockBlob unzippedBlob = container.GetBlockBlobReference(String.Format("{0:yyyy-MM-dd}", lastWriteTime) + " " + Path.GetFileNameWithoutExtension(fileName) + ".csv");
        unzippedBlob.Properties.ContentType = "text/csv";
        using (Stream outputFileStream = await unzippedBlob.OpenWriteAsync())
        {
            await decompressionStream.CopyToAsync(outputFileStream);
            outputFileStream.Flush();
        }
    }
}

5th

using (var inputStream = new MemoryStream())
{
    await zipOutputBlob.DownloadToStreamAsync(inputStream);
    inputStream.Seek(0, SeekOrigin.Begin);

    using (var gzStream = new GZipStream(inputStream, CompressionMode.Decompress))
    {
        using (var outputStream = new MemoryStream())
        {
            gzStream.CopyTo(outputStream);
            byte[] outputBytes = outputStream.ToArray(); // No data. Sad panda. :'(
            string output = Encoding.ASCII.GetString(outputBytes);
            CloudBlockBlob unzippedBlob = container.GetBlockBlobReference(String.Format("{0:yyyy-MM-dd}", lastWriteTime) + " " + Path.GetFileNameWithoutExtension(fileName) + ".csv");
            unzippedBlob.Properties.ContentType = "text/csv";
            await unzippedBlob.UploadTextAsync(output);
        }
    }
}

6th

using (var ms = new MemoryStream())
{
    await zipOutputBlob.DownloadToStreamAsync(ms);
    ms.Seek(0, SeekOrigin.Begin);

    using (DeflateStream decompressionStream = new DeflateStream(ms, CompressionMode.Decompress))
    {
        using (var outputStream = new MemoryStream())
        {
            decompressionStream.CopyTo(outputStream);
            byte[] outputBytes = outputStream.ToArray(); // No data. Sad panda. :'(
            string output = Encoding.ASCII.GetString(outputBytes);
            CloudBlockBlob unzippedBlob = container.GetBlockBlobReference(String.Format("{0:yyyy-MM-dd}", lastWriteTime) + " " + Path.GetFileNameWithoutExtension(fileName) + ".csv");
            unzippedBlob.Properties.ContentType = "text/csv";
            await unzippedBlob.UploadTextAsync(output);
        }
    }
}

Option 5 and 6 fail with this error message on the CopyTo method:

System.Private.CoreLib: Exception while executing function: DefinitiveHealthCare. System.IO.Compression: The archive entry was compressed using an unsupported compression method.

How is this done?

Not sure if this would every be searching via Google, since closed as a duplicate, but I do think my solution would help someone in the future:

private static async Task UnzipDefinitiveFile(string fileName, CloudBlobContainer container, Logger logger, DateTime lastWriteTime, CloudBlockBlob zipOutputBlob)
{
    using (MemoryStream blobMemStream = new MemoryStream())
    {
        await zipOutputBlob.DownloadToStreamAsync(blobMemStream);
        using (ZipArchive archive = new ZipArchive(blobMemStream))
        {
            foreach (ZipArchiveEntry entry in archive.Entries)
            {
                logger.Send(SeverityLevel.Verbose, $"Now processing {entry.FullName}");
                if (entry.FullName != Path.GetFileNameWithoutExtension(fileName) + ".csv")
                    continue;
                string validName = String.Format("{0:yyyy-MM-dd}", lastWriteTime) + " " + Path.GetFileNameWithoutExtension(fileName) + ".csv";
                CloudBlockBlob blockBlob = container.GetBlockBlobReference(validName);
                blockBlob.Properties.ContentType = "text/csv";
                using (var fileStream = entry.Open())
                {
                    await blockBlob.UploadFromStreamAsync(fileStream);
                }
            }
        }
    }
}

The change here is using the ZipArchive class and looping through all the elements in the ZipArchive. In my case there was only one file in the ZipArchive, so I was trying to skip this step which did not work out well.

Also another option could be to just use Azure Data Factory V2. The copy data activity can load a zipped file and it can download a file from an SFTP location.