[GitHub] [spark] cloud-fan commented on a change in pull request #31791: [SPARK-34678][SQL] Add table function registry

[GitBox <gi...@apache.org>]

cloud-fan commented on a change in pull request #31791:
URL: https://github.com/apache/spark/pull/31791#discussion_r594410546



##########
File path: sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/analysis/FunctionRegistry.scala
##########
@@ -707,3 +721,150 @@ object FunctionRegistry {
     (name, (info, outerBuilder))
   }
 }
+
+trait TableFunctionRegistry extends FunctionRegistryBase[LogicalPlan] {
+
+  /** Create a copy of this registry with identical functions as this registry. */
+  override def clone(): TableFunctionRegistry = throw new CloneNotSupportedException()
+}
+
+class SimpleTableFunctionRegistry
+    extends SimpleFunctionRegistryBase[LogicalPlan]
+    with TableFunctionRegistry {
+
+  override def clone(): SimpleTableFunctionRegistry = synchronized {
+    val registry = new SimpleTableFunctionRegistry
+    functionBuilders.iterator.foreach { case (name, (info, builder)) =>
+      registry.registerFunction(name, info, builder)
+    }
+    registry
+  }
+}
+
+object EmptyTableFunctionRegistry
+    extends EmptyFunctionRegistryBase[LogicalPlan]
+    with TableFunctionRegistry {
+
+  override def clone(): TableFunctionRegistry = this
+}
+
+object TableFunctionRegistry {
+
+  type TableFunctionBuilder = Seq[Expression] => LogicalPlan
+
+  /**
+   * A TVF maps argument lists to resolver functions that accept those arguments. Using a map
+   * here allows for function overloading.
+   */
+  private type TVF = Map[ArgumentList, Seq[Any] => LogicalPlan]

Review comment:
       This is different from how the function overload is done for `Expression`. We just add more constructors in the `Expression` sub-classes to accept different parameters list. 




----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

For queries about this service, please contact Infrastructure at:
users@infra.apache.org



---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org