ZipOutputStream Class

This class writes zip files, as defined in the specification for zip files described by PKWare. The compression for this implementation is provided by a managed-code version of Zlib, included with DotNetZip in the classes in the Ionic.Zlib namespace.

This class provides an alternative programming model to the one enabled by the ZipFile class. Use this when creating zip files, as an alternative to the ZipFile class, when you would like to use a Stream type to write the zip file.

Both the ZipOutputStream class and the ZipFile class can be used to create zip files. Both of them support many of the common zip features, including Unicode, different compression levels, and ZIP64. They provide very similar performance when creating zip files.

The ZipFile class is generally easier to use than ZipOutputStream and should be considered a higher-level interface. For example, when creating a zip file via calls to the PutNextEntry() and Write() methods on the ZipOutputStream class, the caller is responsible for opening the file, reading the bytes from the file, writing those bytes into the ZipOutputStream, setting the attributes on the ZipEntry, and setting the created, last modified, and last accessed timestamps on the zip entry. All of these things are done automatically by a call to ZipFile.AddFile(). For this reason, the ZipOutputStream is generally recommended for use only when your application emits arbitrary data, not necessarily data from a filesystem file, directly into a zip file, and does so using a Stream metaphor.

Aside from the differences in programming model, there are other differences in capability between the two classes.

• ZipFile can be used to read and extract zip files, in addition to creating zip files. ZipOutputStream cannot read zip files. If you want to use a stream to read zip files, check out the ZipInputStream class.
• ZipOutputStream does not support the creation of segmented or spanned zip files.
• ZipOutputStream cannot produce a self-extracting archive.

Be aware that the ZipOutputStream class implements the IDisposable interface. In order for ZipOutputStream to produce a valid zip file, you use use it within a using clause (Using in VB), or call the Dispose() method explicitly. See the examples for how to employ a using clause.

Also, a note regarding compression performance: On the desktop .NET Framework, DotNetZip can use a multi-threaded compression implementation that provides significant speed increases on large files, over 300k or so, at the cost of increased memory use at runtime. (The output of the compression is almost exactly the same size). But, the multi-threaded approach incurs a performance hit on smaller files. There's no way for the ZipOutputStream to know whether parallel compression will be beneficial, because the ZipOutputStream does not know how much data you will write through the stream. You may wish to set the ParallelDeflateThreshold property to zero, if you are compressing large files through ZipOutputStream. This will cause parallel compression to be used, always.