[SPARK-5193][SQL] Reconcile Java and Scala UDFRegistration.

1. Removed UDFRegistration as a mixin in SQLContext and made it a field ("udf").
2. For Java UDFs, renamed dataType to returnType, and made that the 2nd argument since the UDF itself can be long.
3. Added all Java UDF registration methods to Scala's UDFRegistration.
4. Documentation

Author: rxin

sql/core/src/main/scala/org/apache/spark/sql/SQLContext.scala

@@ -54,7 +54,6 @@ class SQLContext(@transient val sparkContext: SparkContext)
   extends org.apache.spark.Logging
   with CacheManager
   with ExpressionConversions
-  with UDFRegistration
   with Serializable {
 
   self =>
@@ -338,6 +337,33 @@ class SQLContext(@transient val sparkContext: SparkContext)
    */
   val experimental: ExperimentalMethods = new ExperimentalMethods(this)
 
+  /**
+   * A collection of methods for registering user-defined functions (UDF).
+   *
+   * The following example registers a Scala closure as UDF:
+   * {{{
+   *   sqlContext.udf.register("myUdf", (arg1: Int, arg2: String) => arg2 + arg1)
+   * }}}
+   *
+   * The following example registers a UDF in Java:
+   * {{{
+   *   sqlContext.udf().register("myUDF", DataTypes.StringType,
+   *     new UDF2<Integer, String, String>() {
+   *       @Override
+   *       public String call(Integer arg1, String arg2) {
+   *         return arg2 + arg1;
+   *       }
+   *     });
+   * }}}
+   *
+   * Or, to use Java 8 lambda syntax:
+   * {{{
+   *   sqlContext.udf().register("myUDF", DataTypes.StringType,
+   *     (Integer arg1, String arg2) -> arg2 + arg1));
+   * }}}
+   */
+  val udf: UDFRegistration = new UDFRegistration(this)
+
   protected[sql] class SparkPlanner extends SparkStrategies {
     val sparkContext: SparkContext = self.sparkContext