[SPARK-10419] [SQL] Adding SQLServer support for datetimeoffset types to JdbcDialects

Reading from Microsoft SQL Server over jdbc fails when the table contains datetimeoffset types.

This patch registers a SQLServer JDBC Dialect that maps datetimeoffset to a String, as Microsoft suggest.

Author: Ewan Leith <ewan.leith@realitymine.com>

Closes #8575 from realitymine-coordinator/sqlserver.

sql/core/src/main/scala/org/apache/spark/sql/jdbc/JdbcDialects.scala

@@ -137,6 +137,8 @@ object JdbcDialects {
   registerDialect(MySQLDialect)
   registerDialect(PostgresDialect)
   registerDialect(DB2Dialect)
+  registerDialect(MsSqlServerDialect)
+
 
   /**
    * Fetch the JdbcDialect class corresponding to a given database url.
@@ -260,3 +262,19 @@ case object DB2Dialect extends JdbcDialect {
     case _ => None
   }
 }
+
+/**
+ * :: DeveloperApi ::
+ * Default Microsoft SQL Server dialect, mapping the datetimeoffset types to a String on read.
+ */
+@DeveloperApi
+case object MsSqlServerDialect extends JdbcDialect {
+  override def canHandle(url: String): Boolean = url.startsWith("jdbc:sqlserver")
+  override def getCatalystType(
+      sqlType: Int, typeName: String, size: Int, md: MetadataBuilder): Option[DataType] = {
+    if (typeName.contains("datetimeoffset")) {
+      // String is recommend by Microsoft SQL Server for datetimeoffset types in non-MS clients
+      Some(StringType)
+    } else None
+  }
+}

sql/core/src/test/scala/org/apache/spark/sql/jdbc/JDBCSuite.scala

@@ -408,6 +408,7 @@ class JDBCSuite extends SparkFunSuite with BeforeAndAfter with SharedSQLContext
     assert(JdbcDialects.get("jdbc:mysql://127.0.0.1/db") == MySQLDialect)
     assert(JdbcDialects.get("jdbc:postgresql://127.0.0.1/db") == PostgresDialect)
     assert(JdbcDialects.get("jdbc:db2://127.0.0.1/db") == DB2Dialect)
+    assert(JdbcDialects.get("jdbc:sqlserver://127.0.0.1/db") == MsSqlServerDialect)
     assert(JdbcDialects.get("test.invalid") == NoopDialect)
   }