[SPARK-4901] [SQL] Hot fix for ByteWritables.copyBytes

[chenghao-intel]

HiveInspectors.scala failed in compiling with Hadoop 1, as the BytesWritable.copyBytes is not available in Hadoop 1.

[SparkQA]

Test build #24641 has started for PR 3742 at commit bb04d1f.

• This patch merges cleanly.

[SparkQA]

Test build #24641 has finished for PR 3742 at commit bb04d1f.

• This patch passes all tests.
• This patch merges cleanly.
• This patch adds no public classes.

[AmplabJenkins]

Test PASSed.
Refer to this link for build results (access rights to CI server needed):
https://amplab.cs.berkeley.edu/jenkins//job/SparkPullRequestBuilder/24641/
Test PASSed.

[JoshRosen]

LGTM. It looks like this properly uses getBytes() by also using getLength() to determine which subarray of the getBytes() result contains valid data (see HADOOP-6298).

I suppose this could use Arrays.copyOfRange to cut out one line (like was done in fc616d5#diff-364713d7776956cb8b0a771e9b62f82dR1435), but that's just a minor thing, so I'm going to merge this to fix the Hadoop 1 build. Thanks!