Incorrect Encoding of UTF8 Strings

org.kitesdk.data.hbase.avro.io.ColumnEncoder:writeString should write only from 0 to utf8.getBytesLength instead of the entire UTF8 byte array. To verify this check
MemcmpEncoder.java:writeString.

The reason this is a critical issue is because UTF8 is not immutable and thus the byte[] may have records from a previous deserialization.