fix: prevent native sort crash for Struct(Map(...)) keys

[0lai0]

Which issue does this PR close?

Closes #4123

Rationale for this change

Spark 4.1.1 SQL coverage exposed a case where Comet allows native aggregation for grouping keys that contain nested MapType (for example StructType(MapType(...))).
Although planning succeeds, downstream native sort/row-format handling does not support this shape, so the query fails at execution time instead of falling back to Spark during planning.

This PR makes aggregate planning reject map-containing grouping key types so these queries reliably fall back to Spark.

What changes are included in this PR?

• add a recursive containsMapType helper in CometBaseAggregate
• make aggregate planning fall back when any grouping expression contains MapType directly or nested in StructType/ArrayType
• add regression test in CometAggregateSuite for grouping on Struct(Map(...))
• add regression test in CometAggregateSuite for grouping on Array(Map(...))

How are these changes tested?

• ./mvnw test -Pspark-4.1 -Dtest=none -Dsuites="org.apache.comet.exec.CometAggregateSuite grouping on struct containing map should fallback to Spark"
• ./mvnw test -Pspark-4.1 -Dtest=none -Dsuites="org.apache.comet.exec.CometAggregateSuite group by array of map falls back to Spark (issue #4123)"
• ./mvnw test -Pspark-4.1 -Dtest=none -Dsuites="org.apache.comet.exec.CometAggregateSuite"

[andygrove]

LGTM. Thanks @0lai0

[andygrove]

may need to skip this for Spark 3.x with assume(isSpark40Plus)

[andygrove]

actually isSpark41Plus

[0lai0]

Thanks for point this out. I missed that. I'll add it in next commit.

[andygrove]

@0lai0 test failure on 3.4:

- grouping on struct containing map should fallback to Spark *** FAILED *** (35 milliseconds)
  org.apache.spark.sql.AnalysisException: expression t.col1 cannot be used as a grouping expression because its data type struct<data:map<string,string>> is not an orderable data type.; line 6 pos 9;
Project [col1.data[key]#3959]
+- Sort [col1#3957.data[key] ASC NULLS FIRST], true
   +- Project [col1.data[key]#3959, col1#3957]
      +- Filter isnotnull(col1#3957.data[num])
         +- Aggregate [col1#3957], [col1#3957.data[key] AS col1.data[key]#3959, col1#3957]
            +- SubqueryAlias t
               +- LocalRelation [col1#3957]

[0lai0]

Thanks @andygrove for review. PR updated.

[andygrove]

@0lai0 I think you need to skip more tests with assume(isSpark41Plus, ...) ?

[0lai0]

That sounds like a good idea; I could try it locally. Thanks @andygrove.

[andygrove]

CI is green now @0lai0 🎉

Could you fix the conflict?

[0lai0]

Sure, Thanks @andygrove. PR updated! 🎉