jonvex commented on code in PR #10137:
URL: https://github.com/apache/hudi/pull/10137#discussion_r1408285783
##########
hudi-spark-datasource/hudi-spark-common/src/main/scala/org/apache/spark/sql/execution/datasources/parquet/HoodieFileGroupReaderBasedParquetFileFormat.scala:
##########
@@ -91,73 +95,69 @@ class
HoodieFileGroupReaderBasedParquetFileFormat(tableState: HoodieTableState,
val outputSchema = StructType(requiredSchema.fields ++
partitionSchema.fields)
spark.conf.set("spark.sql.parquet.enableVectorizedReader",
supportBatchResult)
val requiredSchemaWithMandatory =
generateRequiredSchemaWithMandatory(requiredSchema, dataSchema, partitionSchema)
+ val isCount = requiredSchemaWithMandatory.isEmpty
val requiredSchemaSplits = requiredSchemaWithMandatory.fields.partition(f
=> HoodieRecord.HOODIE_META_COLUMNS_WITH_OPERATION.contains(f.name))
val requiredMeta = StructType(requiredSchemaSplits._1)
val requiredWithoutMeta = StructType(requiredSchemaSplits._2)
val augmentedHadoopConf = FSUtils.buildInlineConf(hadoopConf)
- val (baseFileReader, preMergeBaseFileReader, _, _) = buildFileReaders(
+ val (baseFileReader, preMergeBaseFileReader, readerMaps) =
buildFileReaders(
spark, dataSchema, partitionSchema, if (isIncremental)
requiredSchemaWithMandatory else requiredSchema,
filters, options, augmentedHadoopConf, requiredSchemaWithMandatory,
requiredWithoutMeta, requiredMeta)
+
+ val requestedAvroSchema =
AvroConversionUtils.convertStructTypeToAvroSchema(requiredSchema,
sanitizedTableName)
+ val dataAvroSchema =
AvroConversionUtils.convertStructTypeToAvroSchema(dataSchema,
sanitizedTableName)
+
val broadcastedHadoopConf = spark.sparkContext.broadcast(new
SerializableConfiguration(augmentedHadoopConf))
+ val broadcastedDataSchema = spark.sparkContext.broadcast(dataAvroSchema)
+ val broadcastedRequestedSchema =
spark.sparkContext.broadcast(requestedAvroSchema)
val props: TypedProperties = HoodieFileIndex.getConfigProperties(spark,
options)
(file: PartitionedFile) => {
file.partitionValues match {
// Snapshot or incremental queries.
case fileSliceMapping: HoodiePartitionFileSliceMapping =>
val filePath =
sparkAdapter.getSparkPartitionedFileUtils.getPathFromPartitionedFile(file)
- if (FSUtils.isLogFile(filePath)) {
- val partitionValues = fileSliceMapping.getPartitionValues
- val fileSlice =
fileSliceMapping.getSlice(FSUtils.getFileId(filePath.getName).substring(1)).get
- buildFileGroupIterator(
- Option.empty[PartitionedFile => Iterator[InternalRow]],
- partitionValues,
- Option.empty[HoodieBaseFile],
- getLogFilesFromSlice(fileSlice),
- requiredSchemaWithMandatory,
- outputSchema,
- partitionSchema,
- broadcastedHadoopConf.value.value,
- -1,
- -1,
- shouldUseRecordPosition
- )
+ val filegroupName = if (FSUtils.isLogFile(filePath)) {
+ FSUtils.getFileId(filePath.getName).substring(1)
} else {
- fileSliceMapping.getSlice(FSUtils.getFileId(filePath.getName))
match {
- case Some(fileSlice) =>
+ FSUtils.getFileId(filePath.getName)
+ }
+ fileSliceMapping.getSlice(filegroupName) match {
+ case Some(fileSlice) if !isCount =>
+ if (requiredSchema.isEmpty &&
!fileSlice.getLogFiles.findAny().isPresent) {
val hoodieBaseFile = fileSlice.getBaseFile.get()
- val bootstrapFileOpt = hoodieBaseFile.getBootstrapBaseFile
- val partitionValues = fileSliceMapping.getPartitionValues
- val logFiles = getLogFilesFromSlice(fileSlice)
- if (requiredSchemaWithMandatory.isEmpty) {
- val baseFile = createPartitionedFile(partitionValues,
hoodieBaseFile.getHadoopPath, 0, hoodieBaseFile.getFileLen)
- baseFileReader(baseFile)
- } else if (bootstrapFileOpt.isPresent) {
- // TODO: Use FileGroupReader here: HUDI-6942.
- throw new NotImplementedError("Not support reading bootstrap
file")
- } else {
- if (logFiles.isEmpty) {
- throw new IllegalStateException(
- "should not be here since file slice should not have
been broadcasted "
- + "since it has no log or data files")
- }
- buildFileGroupIterator(
- Option(preMergeBaseFileReader),
- partitionValues,
- Option(hoodieBaseFile),
- logFiles,
- requiredSchemaWithMandatory,
- outputSchema,
- partitionSchema,
- broadcastedHadoopConf.value.value,
- 0,
- hoodieBaseFile.getFileLen,
- shouldUseRecordPosition
- )
- }
- // TODO: Use FileGroupReader here: HUDI-6942.
- case _ => baseFileReader(file)
- }
+
baseFileReader(createPartitionedFile(fileSliceMapping.getPartitionValues,
hoodieBaseFile.getHadoopPath, 0, hoodieBaseFile.getFileLen))
+ } else {
+ val readerContext: HoodieReaderContext[InternalRow] = new
SparkFileFormatInternalRowReaderContext(
+ readerMaps)
+ val serializedHadoopConf = broadcastedHadoopConf.value.value
+ val metaClient: HoodieTableMetaClient = HoodieTableMetaClient
+
.builder().setConf(serializedHadoopConf).setBasePath(tableState.tablePath).build
+ val reader = new HoodieFileGroupReader[InternalRow](
+ readerContext,
+ serializedHadoopConf,
+ tableState.tablePath,
+ tableState.latestCommitTimestamp.get,
Review Comment:
good catch, I took that from what was there before and didn't notice
--
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.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
