Compress data streams in Java
Friday, March 02, 2007 09:43 AM
You can compress and decompress any data stream in Java. Peter Mikhalenko shows you just how easy it is to compress data streams with GZIP and Zip data formats.
The classical input/output (I/O) library in Java contains classes that support I/O handling of streams with compressed data. You can easily compress and decompress any text or binary data to or from any I/O stream using either a file or any other stream (e.g., a servlet output stream). In this article, I briefly show how easily you can compress data streams in Java with GZIP and Zip data formats.
Data compressing classes
Data compressing classes use generic I/O streams at a lower
level. It's important that these classes are not a part of symbol streams hierarchy like
Reader and Writer but are based on byte streams InputStream
and OutputStream. This is because the compressing
library works with bytes and not with symbols. Nevertheless, you can always use
mixed streams by converting a byte stream into symbol stream by using InputStreamReader and OutputStreamWriter.
Here are some of the classes that you may use when working with data streams:
- DeflaterOutputStream: This is the base class for all data compressing classes.
- CheckedInputStream: For any input stream InputStream, this class can return a checksum with method getCheckSum().
- CheckedOutputStream: For any output stream OutputStream, this class can return a checksum with method getCheckSum().
- ZipOutputStream: This is a subclass of DeflaterOutputStream; its main purpose is compressing data in Zip format.
- GZIPOutputStream: This is a subclass of DeflaterOutputStream; it compresses data in GZIP format.
- InflaterInputStream: This is a base class for decompressing data.
- ZipInputStream: This subclass of InflaterInputStream can decompress Zip format data.
- GZIPInputStream: This subclass of InflaterInputStream can decompress GZIP format data.
There are a lot of data compressing algorithms, but GZIP and Zip formats are the most frequently used. This is why they are implemented in a standard Java package.
GZIP encoding
GZIP, the simplest compressing method, is ideal for
situations when you have only one data stream that you need to compress. In
Listing A, I compress and then
decompress a file with the help of Java classes located in the java.util.zip package.
To use data compressing classes, you convert them using constructor conversions of I/O classes into the output stream that you need. In Listing A, I used mixed byte and symbol streams; stream uses classes based on the Reader class, while the constructor of GZIPOutputStream uses only streams based on OutputStream but not Writer. This is why when you open a file, the data compressing stream GZIPInputStream is converted into the symbolic stream Reader.
Zip compression
The library for handling Zip data format has much more to
offer than the one for GZIP. It's easy to compress any number of files, and
there is even a special class for reading Zip files. Java uses standard Zip
format, so any compressing or archive utility will be able to read your
compressed data.
Listing B has the
same structure as Listing A, but the number of files
is not limited.
Note the using of classes CheckedOutputStream and CRC32. You can easily get a checksum of your data and control its integrity. Checksum is used for checking that the data has not been changed and/or distorted. CRC32 is a very well-known 32-bit algorithm for checking data integrity.
For each file added into the archive, you must call the method putNextEntry() with the corresponding ZipEntry object. ZipEntry contains everything required for adding into Zip file additional storage information, such as file name, compressed and decompressed sizes, CRC checksum, comments, compressing method, etc. Java does not allow you to set up a password for Zip archive, despite such possibility in the original Zip file format.
Adoption
You can compress and decompress any data stream, which is
not necessarily the case for a file. Data compression is widely used, for
example, in servlet output streams by many servlet applications and application servers (servlet containers) because GZIP-compressed data is also an
Internet standard for transferring data via the HTTP protocol.
Peter V. Mikhalenko is a Sun certified professional who works for Deutsche Bank as a business consultant.
Ccompressed data between client and server
This is useful while reading and writing from and to files. I would like to get the code to transfer compressed data between client and server (TCP socket). Does anyone know a libraray (free or paid) which I can use.
Posted by Vipul Kuruppu on Thursday, March 26 2009 02:41 PM