[GitHub] [spark] cloud-fan commented on a change in pull request #29045: [SPARK-32234][SQL] Spark sql commands are failing on selecting the orc tables

[GitBox]

cloud-fan commented on a change in pull request #29045:
URL: https://github.com/apache/spark/pull/29045#discussion_r454988428



##########
File path: 
sql/core/src/main/scala/org/apache/spark/sql/execution/datasources/v2/orc/OrcPartitionReaderFactory.scala
##########
@@ -66,24 +66,28 @@ case class OrcPartitionReaderFactory(
   override def buildReader(file: PartitionedFile): 
PartitionReader[InternalRow] = {
     val conf = broadcastedConf.value.value
 
-    val resultSchemaString = OrcUtils.orcTypeDescriptionString(resultSchema)
-    OrcConf.MAPRED_INPUT_SCHEMA.setString(conf, resultSchemaString)
     OrcConf.IS_SCHEMA_EVOLUTION_CASE_SENSITIVE.setBoolean(conf, 
isCaseSensitive)
 
     val filePath = new Path(new URI(file.filePath))
 
     val fs = filePath.getFileSystem(conf)
     val readerOptions = OrcFile.readerOptions(conf).filesystem(fs)
-    val requestedColIdsOrEmptyFile =
+    val resultedColPruneInfo =
       Utils.tryWithResource(OrcFile.createReader(filePath, readerOptions)) { 
reader =>
         OrcUtils.requestedColumnIds(
           isCaseSensitive, dataSchema, readDataSchema, reader, conf)
       }
 
-    if (requestedColIdsOrEmptyFile.isEmpty) {
+    if (resultedColPruneInfo.isEmpty) {
       new EmptyPartitionReader[InternalRow]
     } else {
-      val requestedColIds = requestedColIdsOrEmptyFile.get
+      val (requestedColIds, canPruneCols) = resultedColPruneInfo.get
+      val resultSchemaString = if (canPruneCols) {
+        OrcUtils.orcTypeDescriptionString(resultSchema)
+      } else {
+        OrcUtils.orcTypeDescriptionString(StructType(dataSchema.fields ++ 
partitionSchema.fields))
+      }
+      OrcConf.MAPRED_INPUT_SCHEMA.setString(conf, resultSchemaString)

Review comment:
       To avoid duplicated code, we can move these to the `requestedColumnIds` 
method
   ```
   def requestedColumnIds(..., partitionSchema: StructType): Option[Array[Int]] 
= {
     ...
     if (orcFieldNames.isEmpty) {
       None
     } else if (orcFieldNames.forall(_.startsWith("_col"))) {
       OrcConf.MAPRED_INPUT_SCHEMA.setString(conf,
         OrcUtils.orcTypeDescriptionString(StructType(dataSchema.fields ++ 
partitionSchema.fields)))
       ...
     } else {
       OrcConf.MAPRED_INPUT_SCHEMA.setString(conf,
         OrcUtils.orcTypeDescriptionString(StructType(requiredSchema.fields ++ 
partitionSchema.fields)))
       ...
     }
   }
   ```

##########
File path: 
sql/core/src/main/scala/org/apache/spark/sql/execution/datasources/v2/orc/OrcPartitionReaderFactory.scala
##########
@@ -66,24 +66,28 @@ case class OrcPartitionReaderFactory(
   override def buildReader(file: PartitionedFile): 
PartitionReader[InternalRow] = {
     val conf = broadcastedConf.value.value
 
-    val resultSchemaString = OrcUtils.orcTypeDescriptionString(resultSchema)
-    OrcConf.MAPRED_INPUT_SCHEMA.setString(conf, resultSchemaString)
     OrcConf.IS_SCHEMA_EVOLUTION_CASE_SENSITIVE.setBoolean(conf, 
isCaseSensitive)
 
     val filePath = new Path(new URI(file.filePath))
 
     val fs = filePath.getFileSystem(conf)
     val readerOptions = OrcFile.readerOptions(conf).filesystem(fs)
-    val requestedColIdsOrEmptyFile =
+    val resultedColPruneInfo =
       Utils.tryWithResource(OrcFile.createReader(filePath, readerOptions)) { 
reader =>
         OrcUtils.requestedColumnIds(
           isCaseSensitive, dataSchema, readDataSchema, reader, conf)
       }
 
-    if (requestedColIdsOrEmptyFile.isEmpty) {
+    if (resultedColPruneInfo.isEmpty) {
       new EmptyPartitionReader[InternalRow]
     } else {
-      val requestedColIds = requestedColIdsOrEmptyFile.get
+      val (requestedColIds, canPruneCols) = resultedColPruneInfo.get
+      val resultSchemaString = if (canPruneCols) {
+        OrcUtils.orcTypeDescriptionString(resultSchema)
+      } else {
+        OrcUtils.orcTypeDescriptionString(StructType(dataSchema.fields ++ 
partitionSchema.fields))
+      }
+      OrcConf.MAPRED_INPUT_SCHEMA.setString(conf, resultSchemaString)

Review comment:
       To avoid duplicated code, we can move these code to the 
`requestedColumnIds` method
   ```
   def requestedColumnIds(..., partitionSchema: StructType): Option[Array[Int]] 
= {
     ...
     if (orcFieldNames.isEmpty) {
       None
     } else if (orcFieldNames.forall(_.startsWith("_col"))) {
       OrcConf.MAPRED_INPUT_SCHEMA.setString(conf,
         OrcUtils.orcTypeDescriptionString(StructType(dataSchema.fields ++ 
partitionSchema.fields)))
       ...
     } else {
       OrcConf.MAPRED_INPUT_SCHEMA.setString(conf,
         OrcUtils.orcTypeDescriptionString(StructType(requiredSchema.fields ++ 
partitionSchema.fields)))
       ...
     }
   }
   ```




----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

For queries about this service, please contact Infrastructure at:
us...@infra.apache.org



---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org