[SPARK-15321] Fix bug where Array[Timestamp] cannot be encoded/decoded correctly

[smungee]

What changes were proposed in this pull request?

Fix MapObjects.itemAccessorMethod to handle TimestampType. Without this fix, Array[Timestamp] cannot be properly encoded or decoded. To reproduce this, in ExpressionEncoderSuite, if you add the following test case:

encodeDecodeTest(Array(Timestamp.valueOf("2016-01-29 10:00:00")), "array of timestamp")
... you will see that (without this fix) it fails with the following output:

- encode/decode for array of timestamp: [Ljava.sql.Timestamp;@fd9ebde *** FAILED ***
  Exception thrown while decoding
  Converted: [0,1000000010,800000001,52a7ccdc36800]
  Schema: value#61615
  root
  -- value: array (nullable = true)
      |-- element: timestamp (containsNull = true)
  Encoder:
  class[value[0]: array<timestamp>] (ExpressionEncoderSuite.scala:312)

How was this patch tested?

Existing tests

[HyukjinKwon]

(I think it might needs a JIRA because it seems changing existing behaviour)

[smungee]

@HyukjinKwon I think the original author might have just forgotten to add the TimestampType to the match..

[smungee]

The bug is quite easy to reproduce. In ExpressionEncoderSuite, if you add the following test case:

encodeDecodeTest(Array(Timestamp.valueOf("2016-01-29 10:00:00")), "array of timestamp")

... you will see that (without this fix) it fails with the following output:

- encode/decode for array of timestamp: [Ljava.sql.Timestamp;@fd9ebde *** FAILED ***
  Exception thrown while decoding
  Converted: [0,1000000010,800000001,52a7ccdc36800]
  Schema: value#61615
  root
  -- value: array (nullable = true)
      |-- element: timestamp (containsNull = true)


  Encoder:
  class[value[0]: array<timestamp>] (ExpressionEncoderSuite.scala:312)

[techaddict]

@smungee Can you please add tests for this ?

[smungee]

@techaddict Added a test

[rxin]

@smungee sorry to have this gone stale. Do you mind bringing it up to date? This seems useful to fix.

[smungee]

@rxin The code in question seems to have been removed in this commit by @cloud-fan

c36ca65

With the latest SNAPSHOT version, I no longer see the bug. I've updated this PR to include just my additional test-case, just to make sure it doesn't happen again. I'd recommend adding this test-case in.

[SparkQA]

Test build #2999 has finished for PR 13108 at commit 387e6c9.

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

[cloud-fan]

LGTM, merging to master and 2.0, thanks!