[Spark] Make delta.dataSkippingStatsColumns more lenient for nested columns

spark/src/main/scala/org/apache/spark/sql/delta/stats/StatisticsCollection.scala

@@ -459,16 +459,22 @@ object StatisticsCollection extends DeltaCommand {
    * @param name The name of the data skipping column for validating data type.
    * @param dataType The data type of the data skipping column.
    * @param columnPaths The column paths of all valid fields.
+   * @param insideStruct Tracks if we are inside a struct, in which case all types are valid
    */
   private def validateDataSkippingType(
       name: String,
       dataType: DataType,
-      columnPaths: ArrayBuffer[String]): Unit = dataType match {
+      columnPaths: ArrayBuffer[String],
+      insideStruct: Boolean = false): Unit = dataType match {
     case s: StructType =>
       s.foreach { field =>
-        validateDataSkippingType(name + "." + field.name, field.dataType, columnPaths)
+        validateDataSkippingType(name + "." + field.name, field.dataType, columnPaths,
+          insideStruct = true)
       }
     case SkippingEligibleDataType(_) => columnPaths.append(name)
+    case d if insideStruct =>
+      logWarning(s"Data skipping is not supported for column $name of type $dataType")
+      columnPaths.append(name)
     case _ =>
       throw new DeltaIllegalArgumentException(
         errorClass = "DELTA_COLUMN_DATA_SKIPPING_NOT_SUPPORTED_TYPE",

spark/src/test/scala/org/apache/spark/sql/delta/stats/DataSkippingDeltaTests.scala

@@ -599,6 +599,33 @@ trait DataSkippingDeltaTestsBase extends DeltaExcludedBySparkVersionTestMixinShi
     deltaStatsColNamesOpt = Some("b.c")
   )
 
+  testSkipping(
+    "indexed column names - naming a nested column allows nested complex types",
+    """{
+      "a": {
+        "b": [1, 2, 3],
+        "c": [4, 5, 6],
+        "d": 7,
+        "e": 8,
+        "f": {
+          "g": 9
+        }
+      },
+      "i": 10
+    }""".replace("\n", ""),
+    hits = Seq(
+      "i < 0",
+      "a.d > 6",
+      "a.f.g < 10"
+    ),
+    misses = Seq(
+      "a.d < 0",
+      "a.e < 0",
+      "a.f.g < 0"
+    ),
+    deltaStatsColNamesOpt = Some("a")
+  )
+
   testSkipping(
     "indexed column names - index only a subset of leaf columns",
     """{

spark/src/test/scala/org/apache/spark/sql/delta/stats/StatsCollectionSuite.scala

@@ -499,14 +499,13 @@ class StatsCollectionSuite
     ("BINARY", "BinaryType"),
     ("BOOLEAN", "BooleanType"),
     ("ARRAY<TINYINT>", "ArrayType(ByteType,true)"),
-    ("MAP<DATE, INT>", "MapType(DateType,IntegerType,true)"),
-    ("STRUCT<c60:INT, c61:ARRAY<INT>>", "ArrayType(IntegerType,true)")
+    ("MAP<DATE, INT>", "MapType(DateType,IntegerType,true)")
   ).foreach { case (invalidType, typename) =>
     val tableName1 = "delta_table_1"
     val tableName2 = "delta_table_2"
     test(s"Delta statistic column: invalid data type $invalidType") {
       withTable(tableName1, tableName2) {
-        val columnName = if (typename.equals("ArrayType(IntegerType,true)")) "c2.c61" else "c2"
+        val columnName = "c2"
         val exceptOne = intercept[DeltaIllegalArgumentException] {
           sql(
             s"create table $tableName1 (c1 long, c2 $invalidType) using delta " +
@@ -530,7 +529,7 @@ class StatsCollectionSuite
 
     test(s"Delta statistic column: invalid data type $invalidType in nested column") {
       withTable(tableName1, tableName2) {
-        val columnName = if (typename == "ArrayType(IntegerType,true)") "c2.c21.c61" else "c2.c21"
+        val columnName = "c2.c21"
         val exceptOne = intercept[DeltaIllegalArgumentException] {
           sql(
             s"create table $tableName1 (c1 long, c2 STRUCT<c20:INT, c21:$invalidType>) " +
@@ -541,16 +540,6 @@ class StatsCollectionSuite
           exceptOne.getErrorClass == "DELTA_COLUMN_DATA_SKIPPING_NOT_SUPPORTED_TYPE" &&
             exceptOne.getMessageParametersArray.toSeq == Seq(columnName, typename)
         )
-        val exceptTwo = intercept[DeltaIllegalArgumentException] {
-          sql(
-            s"create table $tableName1 (c1 long, c2 STRUCT<c20:INT, c21:$invalidType>) " +
-              s"using delta TBLPROPERTIES('delta.dataSkippingStatsColumns' = 'c2')"
-          )
-        }
-        assert(
-          exceptTwo.getErrorClass == "DELTA_COLUMN_DATA_SKIPPING_NOT_SUPPORTED_TYPE" &&
-          exceptTwo.getMessageParametersArray.toSeq == Seq(columnName, typename)
-        )
         sql(s"create table $tableName2 (c1 long, c2 STRUCT<c20:INT, c21:$invalidType>) using delta")
         val exceptThree = interceptWithUnwrapping[DeltaIllegalArgumentException] {
           sql(
@@ -561,13 +550,6 @@ class StatsCollectionSuite
           exceptThree.getErrorClass == "DELTA_COLUMN_DATA_SKIPPING_NOT_SUPPORTED_TYPE" &&
           exceptThree.getMessageParametersArray.toSeq == Seq(columnName, typename)
         )
-        val exceptFour = interceptWithUnwrapping[DeltaIllegalArgumentException] {
-          sql(s"ALTER TABLE $tableName2 SET TBLPROPERTIES('delta.dataSkippingStatsColumns'='c2')")
-        }
-        assert(
-          exceptFour.getErrorClass == "DELTA_COLUMN_DATA_SKIPPING_NOT_SUPPORTED_TYPE" &&
-          exceptFour.getMessageParametersArray.toSeq == Seq(columnName, typename)
-        )
       }
     }
   }
@@ -608,7 +590,8 @@ class StatsCollectionSuite
 
   Seq(
     "BIGINT", "DATE", "DECIMAL(3, 2)", "DOUBLE", "FLOAT", "INT", "SMALLINT", "STRING",
-    "TIMESTAMP", "TIMESTAMP_NTZ", "TINYINT"
+    "TIMESTAMP", "TIMESTAMP_NTZ", "TINYINT", "STRUCT<c3: BIGINT>",
+    "STRUCT<c3: BIGINT, c4: ARRAY<BIGINT>>"
   ).foreach { validType =>
     val tableName1 = "delta_table_1"
     val tableName2 = "delta_table_2"