[BUG] `getColumnTypeName` for timestamp depends on cursor on/off

What is the bug?

getColumnTypeName for timestamp depends whether cursor feature is active

How can one reproduce the bug?

Steps to reproduce the behavior:

1. Create an index and a doc:

curl -s -H 'Content-Type: application/json' -XPUT "http://localhost:9200/date_type_index?pretty" -d '{"mappings": {"properties": {"release_date": {"type": "date" }}}}'
curl -s -H 'Content-Type: application/json' -XPOST "http://localhost:9200/date_type_index/_doc?pretty" -d '{"release_date": "2022-11-16"}'

2. Query it without cursor:

Connection conn = driver.connect(connStr, properties);
Statement stmt = conn.createStatement(ResultSet.TYPE_FORWARD_ONLY, ResultSet.CONCUR_READ_ONLY);
ResultSet rs = stmt.executeQuery("select * from date_type_index");
ResultSetMetaData rsmd = rs.getMetaData();
String typeName = rsmd.getColumnTypeName("release_date");

3. typeName is TIMESTAMP
4. Query it with cursor:

Connection conn = driver.connect(connStr, properties);
Statement stmt = conn.createStatement(ResultSet.TYPE_FORWARD_ONLY, ResultSet.CONCUR_READ_ONLY);
stmt.setFetchSize(5);
ResultSet rs = stmt.executeQuery("select * from date_type_index");
ResultSetMetaData rsmd = rs.getMetaData();
String typeName = rsmd.getColumnTypeName("release_date");

5. typeName is DATE

What is the expected behavior?

It should be always timestamp since SQL plugin reports this type:

{
  "schema": [
    {
      "name": "release_date",
      "type": "timestamp"
    }
  ],
  "datarows": [
    [
      "2022-11-16 00:00:00"
    ]
  ],
  "total": 1,
  "size": 1,
  "status": 200
}

What is your host/environment?

• OpenSearch 2.5
• JDBC 5e8d914d @ main

Do you have any screenshots?

N/A

Do you have any additional context?

Thanks @akuzin1 for reporting that

Start at the JDBC ResultSetMetaData.getColumnTypeName entry point and reproduce the issue with the reported cursor-off and cursor-on statements. Compare the metadata for release_date in both cases; done when both paths report TIMESTAMP consistently with the SQL plugin schema.