Improve SnappyCompressor buffer expansion performance

The existing implementation of SnappyCompressor will only allocate enough bytes for the buffer passed into setInput().  This leads to suboptimal performance when there are patterns of writes that cause repeated buffer expansions.  In the worst case it must copy the entire buffer for every single invocation of setInput()

Instead of allocating a buffer of size current + write length,  there should be an expansion strategy that reduces the amount of copying required.

**Reporter**: [Andrew Baranec](https://issues.apache.org/jira/secure/ViewProfile.jspa?name=abaranec)

_{**Note**: *This issue was originally created as [PARQUET-2184](https://issues.apache.org/jira/browse/PARQUET-2184). Please see the [migration documentation](https://issues.apache.org/jira/browse/PARQUET-2502) for further details.*}

Locate SnappyCompressor and its setInput() implementation, then inspect how buffer allocation and copying behave across repeated writes. The work is done when buffer expansion avoids copying the entire buffer on every invocation while preserving the compressor's existing input behavior.

索引モデルが issue の本文から書いたものです。