The uncompressed data are retrieved from compressed streams
during most operations, but the streams are still compressed. The
uncompressed data are sometimes cached so that later operations
(including the actual decompression of the streams) need not invoke
the decompression algorithm.
This increases the memory usage, and it is not desirable if the
document is to be kept in memory for an extended period of
This method discards the cached, decompressed data so that a
re-read of the document is not required to release the memory.
You can apply the effect to multiple stream objects using