jackylk commented on a change in pull request #3488: [CARBONDATA-3599] Support
auto ingest data written by Flink
URL: https://github.com/apache/carbondata/pull/3488#discussion_r360407178
##########
File path:
integration/spark-common/src/main/scala/org/apache/carbondata/spark/load/DataLoadProcessBuilderOnSpark.scala
##########
@@ -369,6 +378,72 @@ object DataLoadProcessBuilderOnSpark {
carbonTaskInfo.setTaskId(CarbonUtil.generateUUID())
ThreadLocalTaskInfo.setCarbonTaskInfo(carbonTaskInfo)
}
+
+ /**
+ * create CarbonLoadModel for global_sort
+ */
+ def createLoadModelForGlobalSort(
+ sparkSession: SparkSession,
+ carbonTable: CarbonTable
+ ): CarbonLoadModel = {
+ val conf = SparkSQLUtil.sessionState(sparkSession).newHadoopConf()
+ CarbonTableOutputFormat.setDatabaseName(conf, carbonTable.getDatabaseName)
+ CarbonTableOutputFormat.setTableName(conf, carbonTable.getTableName)
+ CarbonTableOutputFormat.setCarbonTable(conf, carbonTable)
+ val fieldList = carbonTable.getCreateOrderColumn
+ .asScala
+ .map { column =>
+ new StructField(column.getColName, column.getDataType)
+ }
+ CarbonTableOutputFormat.setInputSchema(conf, new
StructType(fieldList.asJava))
+ val loadModel = CarbonTableOutputFormat.getLoadModel(conf)
+ loadModel.setSerializationNullFormat(
+ TableOptionConstant.SERIALIZATION_NULL_FORMAT.getName + ",\\N")
+ loadModel.setBadRecordsLoggerEnable(
+ TableOptionConstant.BAD_RECORDS_LOGGER_ENABLE.getName + ",false")
+ loadModel.setBadRecordsAction(
+ TableOptionConstant.BAD_RECORDS_ACTION.getName + ",force")
+ loadModel.setIsEmptyDataBadRecord(
+ DataLoadProcessorConstants.IS_EMPTY_DATA_BAD_RECORD + ",false")
+ val globalSortPartitions =
+
carbonTable.getTableInfo.getFactTable.getTableProperties.get("global_sort_partitions")
+ if (globalSortPartitions != null) {
+ loadModel.setGlobalSortPartitions(globalSortPartitions)
+ }
+ loadModel
+ }
+
+ /**
+ * create DataFrame basing on specified splits
+ */
+ def createInputDataFrame(
+ sparkSession: SparkSession,
+ carbonTable: CarbonTable,
+ splits: Seq[InputSplit]
+ ): DataFrame = {
+ val columns = carbonTable
+ .getCreateOrderColumn
+ .asScala
+ .map(_.getColName)
+ .toArray
+ val schema = SparkTypeConverter.createSparkSchema(carbonTable, columns)
+ val rdd: RDD[Row] = new CarbonScanRDD[CarbonRow](
+ sparkSession,
+ columnProjection = new CarbonProjection(columns),
+ null,
+ carbonTable.getAbsoluteTableIdentifier,
+ carbonTable.getTableInfo.serialize,
+ carbonTable.getTableInfo,
+ new CarbonInputMetrics,
+ null,
+ null,
+ classOf[CarbonRowReadSupport],
+ splits.asJava)
+ .map { row =>
+ new GenericRow(row.getData.asInstanceOf[Array[Any]])
+ }
+ sparkSession.createDataFrame(rdd, schema)
Review comment:
Yes. Since this is fixed in #3515, I suggest first merge this PR then fix it
in #3515
----------------------------------------------------------------
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
With regards,
Apache Git Services
