feat: do not fallback to Spark for distincts

Author: comphead

spark/src/test/scala/org/apache/comet/CometFuzzAggregateSuite.scala

@@ -79,7 +79,7 @@ class CometFuzzAggregateSuite extends CometFuzzTestBase {
     }
   }
 
-  // Not yet supported
+  // COUNT(distinct x, y, z, ...) not yet supported
   // https://github.com/apache/datafusion-comet/issues/2292
   test("count distinct multiple values and group by multiple column") {
     val df = spark.read.parquet(filename)

spark/src/test/scala/org/apache/comet/exec/CometExecSuite.scala

@@ -1031,6 +1031,8 @@ class CometExecSuite extends CometTestBase {
           |GROUP BY key
               """.stripMargin)
 
+      // The above query uses SUM(DISTINCT) and count(distinct value1, value2)
+      // which is not yet supported
       checkSparkAnswer(df)
       val subPlan = stripAQEPlan(df.queryExecution.executedPlan).collectFirst {
         case s: CometHashAggregateExec => s