[SPARK-31116][SQL] Fix nested schema case-sensitivity in ParquetRowConverter

sql/core/src/main/scala/org/apache/spark/sql/execution/datasources/parquet/ParquetReadSupport.scala

@@ -122,12 +122,15 @@ class ParquetReadSupport(val convertTz: Option[ZoneId],
       keyValueMetaData: JMap[String, String],
       fileSchema: MessageType,
       readContext: ReadContext): RecordMaterializer[InternalRow] = {
+    val caseSensitive = conf.getBoolean(SQLConf.CASE_SENSITIVE.key,
+      SQLConf.CASE_SENSITIVE.defaultValue.get)
     val parquetRequestedSchema = readContext.getRequestedSchema
     new ParquetRecordMaterializer(
       parquetRequestedSchema,
       ParquetReadSupport.expandUDT(catalystRequestedSchema),
       new ParquetToSparkSchemaConverter(conf),
-      convertTz)
+      convertTz,
+      caseSensitive)
   }
 }
 

sql/core/src/main/scala/org/apache/spark/sql/execution/datasources/parquet/ParquetRecordMaterializer.scala

@@ -36,11 +36,18 @@ private[parquet] class ParquetRecordMaterializer(
     parquetSchema: MessageType,
     catalystSchema: StructType,
     schemaConverter: ParquetToSparkSchemaConverter,
-    convertTz: Option[ZoneId])
+    convertTz: Option[ZoneId],
+    caseSensitive: Boolean)
   extends RecordMaterializer[InternalRow] {
 
   private val rootConverter =
-    new ParquetRowConverter(schemaConverter, parquetSchema, catalystSchema, convertTz, NoopUpdater)
+    new ParquetRowConverter(
+      schemaConverter,
+      parquetSchema,
+      catalystSchema,
+      convertTz,
+      caseSensitive,
+      NoopUpdater)
 
   override def getCurrentRecord: InternalRow = rootConverter.currentRecord
 

sql/core/src/main/scala/org/apache/spark/sql/execution/datasources/parquet/ParquetRowConverter.scala

@@ -33,7 +33,7 @@ import org.apache.parquet.schema.PrimitiveType.PrimitiveTypeName.{BINARY, DOUBLE
 import org.apache.spark.internal.Logging
 import org.apache.spark.sql.catalyst.InternalRow
 import org.apache.spark.sql.catalyst.expressions._
-import org.apache.spark.sql.catalyst.util.{ArrayBasedMapData, DateTimeUtils, GenericArrayData}
+import org.apache.spark.sql.catalyst.util.{ArrayBasedMapData, CaseInsensitiveMap, DateTimeUtils, GenericArrayData}
 import org.apache.spark.sql.catalyst.util.DateTimeUtils.SQLTimestamp
 import org.apache.spark.sql.types._
 import org.apache.spark.unsafe.types.UTF8String
@@ -126,6 +126,7 @@ private[parquet] class ParquetRowConverter(
     parquetType: GroupType,
     catalystType: StructType,
     convertTz: Option[ZoneId],
+    caseSensitive: Boolean,
     updater: ParentContainerUpdater)
   extends ParquetGroupConverter(updater) with Logging {
 
@@ -176,10 +177,22 @@ private[parquet] class ParquetRowConverter(
    */
   def currentRecord: InternalRow = currentRow
 
+
   // Converters for each field.
   private[this] val fieldConverters: Array[Converter with HasParentContainerUpdater] = {
+    // For letter case issue, create name to field index based on case sensitivity
+    val catalystFieldNameToIndex = if (caseSensitive) {
+      catalystType.fieldNames.zipWithIndex.toMap
+    } else {
+      CaseInsensitiveMap(catalystType.fieldNames.zipWithIndex.toMap)
+    }
+
     parquetType.getFields.asScala.map { parquetField =>
-      val fieldIndex = catalystType.fieldIndex(parquetField.getName)
+      val fieldIndex = catalystFieldNameToIndex.getOrElse(parquetField.getName,
+        throw new IllegalArgumentException(
+          s"${parquetField.getName} does not exist. " +
+            s"Available: ${catalystType.fieldNames.mkString(", ")}")
+      )
       val catalystField = catalystType(fieldIndex)
       // Converted field value should be set to the `fieldIndex`-th cell of `currentRow`
       newConverter(parquetField, catalystField.dataType, new RowUpdater(currentRow, fieldIndex))
@@ -348,7 +361,7 @@ private[parquet] class ParquetRowConverter(
           }
         }
         new ParquetRowConverter(
-          schemaConverter, parquetType.asGroupType(), t, convertTz, wrappedUpdater)
+          schemaConverter, parquetType.asGroupType(), t, convertTz, caseSensitive, wrappedUpdater)
 
       case t =>
         throw new RuntimeException(

sql/core/src/test/scala/org/apache/spark/sql/execution/datasources/parquet/ParquetQuerySuite.scala

@@ -804,6 +804,162 @@ abstract class ParquetQuerySuite extends QueryTest with ParquetTest with SharedS
     }
   }
 
+  test("SPARK-31116: Select simple parquet columns correctly in case insensitive manner") {
+    withSQLConf(SQLConf.CASE_SENSITIVE.key -> "false") {
+      withTempPath { dir => {
+        val path = dir.getCanonicalPath
+
+        // Prepare values for testing specific parquet record reader
+        Seq("A").toDF("camelCase").write.parquet(path)
+
+        val exactSchema = new StructType().add("camelCase", StringType)
+        checkAnswer(spark.read.schema(exactSchema).parquet(path), Row("A"))
+
+        // In case insensitive manner, parquet's column cases are ignored
+        val caseInsensitiveSchema = new StructType().add("camelcase", StringType)
+        checkAnswer(spark.read.schema(caseInsensitiveSchema).parquet(path), Row("A"))
+      }}
+    }
+  }
+
+  test("SPARK-31116: Select nested parquet columns correctly in case insensitive manner") {
+    withSQLConf(SQLConf.CASE_SENSITIVE.key -> "false") {
+      withTempPath { dir => {
+        val path = dir.getCanonicalPath
+
+        // Prepare values for testing nested parquet data
+        spark
+          .range(1L)
+          .selectExpr("NAMED_STRUCT('lowercase', id, 'camelCase', id + 1) AS StructColumn")
+          .write.parquet(path)
+
+        val exactSchema = new StructType()
+          .add(
+            "StructColumn",
+            new StructType()
+              .add("lowercase", LongType)
+              .add("camelCase", LongType))
+        checkAnswer(spark.read.schema(exactSchema).parquet(path), Row(Row(0, 1)))
+
+        // In case insensitive manner, parquet's column cases are ignored
+        val innerColumnCaseInsensitiveSchema = new StructType()
+          .add(
+            "StructColumn",
+            new StructType()
+              .add("LowerCase", LongType)
+              .add("camelcase", LongType))
+        checkAnswer(
+          spark.read.schema(innerColumnCaseInsensitiveSchema).parquet(path),
+          Row(Row(0, 1)))
+
+        val rootColumnCaseInsensitiveSchema = new StructType()
+          .add(
+            "structColumn",
+            new StructType()
+              .add("lowercase", LongType)
+              .add("camelCase", LongType))
+        checkAnswer(
+          spark.read.schema(rootColumnCaseInsensitiveSchema).parquet(path),
+          Row(Row(0, 1)))
+      }}
+    }
+  }
+
+  test("SPARK-31116: Select simple parquet columns correctly in case sensitive manner") {
+    withSQLConf(SQLConf.CASE_SENSITIVE.key -> "true") {
+      withTempPath { dir => {
+        val path = dir.getCanonicalPath
+
+        // Prepare values for testing specific parquet record reader
+        Seq("A").toDF("camelCase").write.parquet(path)
+
+        val exactSchema = new StructType().add("camelCase", StringType)
+        checkAnswer(spark.read.schema(exactSchema).parquet(path), Row("A"))
+
+        // In case sensitive manner, different letter case does not read column
+        val caseInsensitiveSchema = new StructType().add("camelcase", StringType)
+        checkAnswer(spark.read.schema(caseInsensitiveSchema).parquet(path), Row(null))
+
+        // It also properly work in combined schema
+        val combinedSchema = new StructType()
+          .add("camelCase", StringType)
+          .add("camelcase", StringType)
+        checkAnswer(spark.read.schema(combinedSchema).parquet(path), Row("A", null))
+      }}
+    }
+  }
+
+  test("SPARK-31116: Select nested parquet columns correctly in case sensitive manner") {
+    withSQLConf(SQLConf.CASE_SENSITIVE.key -> "true") {
+      withTempPath { dir => {
+        val path = dir.getCanonicalPath
+
+        // Prepare values for testing nested parquet data
+        spark
+          .range(1)
+          .selectExpr("NAMED_STRUCT('lowercase', id, 'camelCase', id + 1) AS StructColumn")
+          .write.parquet(path)
+
+        val exactSchema = new StructType()
+          .add(
+            "StructColumn",
+            new StructType()
+              .add("lowercase", LongType)
+              .add("camelCase", LongType))
+        checkAnswer(spark.read.schema(exactSchema).parquet(path), Row(Row(0, 1)))
+
+        val innerColumnCaseInsensitiveSchema = new StructType()
+          .add(
+            "StructColumn",
+            new StructType()
+              .add("LowerCase", LongType)
+              .add("camelcase", LongType))
+        checkAnswer(
+          spark.read.schema(innerColumnCaseInsensitiveSchema).parquet(path),
+          Row(null))
+
+        val innerPartialColumnCaseInsensitiveSchema = new StructType()
+          .add(
+            "StructColumn",
+            new StructType()
+              .add("lowercase", LongType)
+              .add("camelcase", LongType))
+        checkAnswer(
+          spark.read.schema(innerPartialColumnCaseInsensitiveSchema).parquet(path),
+          Row(Row(0, null)))
+
+        val rootColumnCaseInsensitiveSchema = new StructType()
+          .add(
+            "structColumn",
+            new StructType()
+              .add("lowercase", LongType)
+              .add("camelCase", LongType))
+        checkAnswer(
+          spark.read.schema(rootColumnCaseInsensitiveSchema).parquet(path),
+          Row(null))
+
+        val combinedSchema = new StructType()
+          .add(
+            "StructColumn",
+            new StructType()
+              .add("lowercase", LongType)
+              .add("camelCase", LongType)
+              .add("LowerCase", LongType)
+              .add("camelcase", LongType))
+          .add(
+            "structColumn",
+            new StructType()
+              .add("lowercase", LongType)
+              .add("camelCase", LongType)
+              .add("LowerCase", LongType)
+              .add("camelcase", LongType))
+        checkAnswer(
+          spark.read.schema(combinedSchema).parquet(path),
+          Row(Row(0, 1, null, null), null))
+      }}
+    }
+  }
+
   test("Migration from INT96 to TIMESTAMP_MICROS timestamp type") {
     def testMigration(fromTsType: String, toTsType: String): Unit = {
       def checkAppend(write: DataFrameWriter[_] => Unit, readback: => DataFrame): Unit = {