lanyuanxiaoyao/hudi
1
0
Fork 0
Code Activity

[HUDI-1105] Adding dedup support for Bulk Insert w/ Rows (#2206)

Browse Source
This commit is contained in:
Sivabalan Narayanan
2021-07-07 17:38:26 -04:00
committed by GitHub
parent 8f7ad8b178
commit 16e90d30ea
7 changed files with 265 additions and 23 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
hudi-spark-datasource/hudi-spark/src/main/scala/org/apache/hudi/HoodieSparkSqlWriter.scala
View File

@@ -38,6 +38,7 @@ import org.apache.hudi.exception.HoodieException
import org.apache.hudi.execution.bulkinsert.BulkInsertInternalPartitionerWithRowsFactory
import org.apache.hudi.hive.util.ConfigUtils
import org.apache.hudi.hive.{HiveSyncConfig, HiveSyncTool}
import org.apache.hudi.index.SparkHoodieIndex
import org.apache.hudi.internal.DataSourceInternalWriterHelper
import org.apache.hudi.keygen.factory.HoodieSparkKeyGeneratorFactory
import org.apache.hudi.sync.common.AbstractSyncTool
@@ -345,8 +346,9 @@ object HoodieSparkSqlWriter {
}
val arePartitionRecordsSorted = bulkInsertPartitionerRows.arePartitionRecordsSorted();
parameters.updated(HoodieInternalConfig.BULKINSERT_ARE_PARTITIONER_RECORDS_SORTED, arePartitionRecordsSorted.toString)
val isGlobalIndex = SparkHoodieIndex.isGlobalIndex(writeConfig)
val hoodieDF = HoodieDatasetBulkInsertHelper.prepareHoodieDatasetForBulkInsert(sqlContext, writeConfig, df, structName, nameSpace,
bulkInsertPartitionerRows)
bulkInsertPartitionerRows, isGlobalIndex)
if (SPARK_VERSION.startsWith("2.")) {
hoodieDF.write.format("org.apache.hudi.internal")
.option(DataSourceInternalWriterHelper.INSTANT_TIME_OPT_KEY, instantTime)