[SPARK-34180][SQL] Revert SPARK-33888

[sarutak]

What changes were proposed in this pull request?

This PR reverts SPARK-33888 (#30902) .
That PR has the two serious problems.

1. A regression which affects PostgresDialect causing PostgreSQLIntegrationSuite to fail.
2. That PR changes the JDBC-Catalyst type mapping. sql.Types.TIME is mapped to sql.types.IntegerType so we'll get millisecond value. But the start time point is different between MsSqlServer and other RDBMSs. For MsSqlServer, it starts from 1900-01-01 00:00:00 while 1970-01-01 00:00:00 for other RDBMS.

More about problem-2, the following assertion passes before SPARK-33888.

// This is from jdbc.MsSqlServerIntegrationSuite
assert(row.getAs[Timestamp](5).equals(Timestamp.valueOf("1900-01-01 13:31:24.0")))

But after SPARK-33888, the following modified assertion fails because the millisecond value is considered starting from 1970-01-01 00:00:00.

assert(row.getAs[Integer](5).equals(Timestamp.valueOf("1900-01-01 13:31:24.0").getTime.toInt))

Why are the changes needed?

To fix the serious issues.

Does this PR introduce any user-facing change?

No.

How was this patch tested?

This is just reverting a previous change.

[cloud-fan]

What do you mean by "start time point"? The TIME type has no date part and should return seconds(with certain precision) of the hour:minute:second fields.

[sarutak]

@cloud-fan

What do you mean by "start time point"? The TIME type has no date part and should return seconds(with certain precision) of the hour:minute:second fields.

Yes it has no date part. So in MsSqlServer, the date part is assumed 1900-01-01 00:00:00 while other RDBMS assumes it as 1970-01-01 00:00:00.
https://docs.microsoft.com/ja-jp/sql/t-sql/data-types/time-transact-sql?view=sql-server-ver15

[sarutak]

@cloud-fan
Or, do you think the assert(row.getAs[Timestamp](5).equals(Timestamp.valueOf("1900-01-01 13:31:24.0"))) is wrong because it contains date part?

[SparkQA]

Kubernetes integration test starting
URL: https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder-K8s/38890/

[cloud-fan]

@sarutak can you take a look at https://github.com/apache/spark/pull/30902/files#r561601546 ? I think the test is wrong.

[sarutak]

@cloud-fan O.K, I'll close this PR and open another PR to fix the wrong assertion. Thanks.

[SparkQA]

Kubernetes integration test status success
URL: https://amplab.cs.berkeley.edu/jenkins/job/SparkPullRequestBuilder-K8s/38890/