[SPARK-30339][SQL][2.4] Avoid to fail twice in function lookup

### What changes were proposed in this pull request?

Backported from [pr#26994](#26994). Currently if function lookup fails, spark will give it a second chance by casting decimal type to double type. But for cases where decimal type doesn't exist, it's meaningless to lookup again and cause extra cost like unnecessary metastore access. We should throw exceptions directly in these cases.

### Does this PR introduce any user-facing change?

No.

### How was this patch tested?

Covered by existing tests.

Closes #27054 from wzhfy/avoid_udf_fail_twice-2.4.

Authored-by: Zhenhua Wang <wzh_zju@163.com>
Signed-off-by: Zhenhua Wang <wzh_zju@163.com>

sql/hive/src/main/scala/org/apache/spark/sql/hive/HiveSessionCatalog.scala

@@ -116,7 +116,7 @@ private[sql] class HiveSessionCatalog(
     try {
       lookupFunction0(name, children)
     } catch {
-      case NonFatal(_) =>
+      case NonFatal(_) if children.exists(_.dataType.isInstanceOf[DecimalType]) =>
         // SPARK-16228 ExternalCatalog may recognize `double`-type only.
         val newChildren = children.map { child =>
           if (child.dataType.isInstanceOf[DecimalType]) Cast(child, DoubleType) else child

sql/hive/src/test/scala/org/apache/spark/sql/hive/execution/HiveUDFSuite.scala

@@ -148,13 +148,6 @@ class HiveUDFSuite extends QueryTest with TestHiveSingleton with SQLTestUtils {
       sql("SELECT array(max(key), max(key)) FROM src").collect().toSeq)
   }
 
-  test("SPARK-16228 Percentile needs explicit cast to double") {
-    sql("select percentile(value, cast(0.5 as double)) from values 1,2,3 T(value)")
-    sql("select percentile_approx(value, cast(0.5 as double)) from values 1.0,2.0,3.0 T(value)")
-    sql("select percentile(value, 0.5) from values 1,2,3 T(value)")
-    sql("select percentile_approx(value, 0.5) from values 1.0,2.0,3.0 T(value)")
-  }
-
   test("Generic UDAF aggregates") {
 
     checkAnswer(sql(