1
0

[HUDI-1104] Adding support for UserDefinedPartitioners and SortModes to BulkInsert with Rows (#3149)

This commit is contained in:
Sivabalan Narayanan
2021-07-07 11:15:25 -04:00
committed by GitHub
parent 55ecbc662e
commit ea9e5d0e8b
31 changed files with 618 additions and 82 deletions

View File

@@ -19,6 +19,7 @@
package org.apache.hudi.spark3.internal;
import org.apache.hudi.DataSourceUtils;
import org.apache.hudi.config.HoodieInternalConfig;
import org.apache.hudi.config.HoodieWriteConfig;
import org.apache.hudi.internal.BaseDefaultSource;
import org.apache.hudi.internal.DataSourceInternalWriterHelper;
@@ -47,9 +48,11 @@ public class DefaultSource extends BaseDefaultSource implements TableProvider {
String instantTime = properties.get(DataSourceInternalWriterHelper.INSTANT_TIME_OPT_KEY);
String path = properties.get("path");
String tblName = properties.get(HoodieWriteConfig.TABLE_NAME.key());
boolean arePartitionRecordsSorted = Boolean.parseBoolean(properties.getOrDefault(HoodieInternalConfig.BULKINSERT_ARE_PARTITIONER_RECORDS_SORTED,
Boolean.toString(HoodieInternalConfig.DEFAULT_BULKINSERT_ARE_PARTITIONER_RECORDS_SORTED)));
// 1st arg to createHooodieConfig is not really reuqired to be set. but passing it anyways.
HoodieWriteConfig config = DataSourceUtils.createHoodieConfig(properties.get(HoodieWriteConfig.AVRO_SCHEMA.key()), path, tblName, properties);
return new HoodieDataSourceInternalTable(instantTime, config, schema, getSparkSession(),
getConfiguration());
getConfiguration(), arePartitionRecordsSorted);
}
}

View File

@@ -37,9 +37,9 @@ public class HoodieBulkInsertDataInternalWriter implements DataWriter<InternalRo
private final BulkInsertDataInternalWriterHelper bulkInsertWriterHelper;
public HoodieBulkInsertDataInternalWriter(HoodieTable hoodieTable, HoodieWriteConfig writeConfig,
String instantTime, int taskPartitionId, long taskId, StructType structType) {
String instantTime, int taskPartitionId, long taskId, StructType structType, boolean arePartitionRecordsSorted) {
this.bulkInsertWriterHelper = new BulkInsertDataInternalWriterHelper(hoodieTable,
writeConfig, instantTime, taskPartitionId, taskId, 0, structType);
writeConfig, instantTime, taskPartitionId, taskId, 0, structType, arePartitionRecordsSorted);
}
@Override

View File

@@ -35,18 +35,20 @@ public class HoodieBulkInsertDataInternalWriterFactory implements DataWriterFact
private final HoodieTable hoodieTable;
private final HoodieWriteConfig writeConfig;
private final StructType structType;
private final boolean arePartitionRecordsSorted;
public HoodieBulkInsertDataInternalWriterFactory(HoodieTable hoodieTable, HoodieWriteConfig writeConfig,
String instantTime, StructType structType) {
String instantTime, StructType structType, boolean arePartitionRecordsSorted) {
this.hoodieTable = hoodieTable;
this.writeConfig = writeConfig;
this.instantTime = instantTime;
this.structType = structType;
this.arePartitionRecordsSorted = arePartitionRecordsSorted;
}
@Override
public DataWriter<InternalRow> createWriter(int partitionId, long taskId) {
return new HoodieBulkInsertDataInternalWriter(hoodieTable, writeConfig, instantTime, partitionId, taskId,
structType);
structType, arePartitionRecordsSorted);
}
}

View File

@@ -45,13 +45,15 @@ public class HoodieDataSourceInternalBatchWrite implements BatchWrite {
private final String instantTime;
private final HoodieWriteConfig writeConfig;
private final StructType structType;
private final boolean arePartitionRecordsSorted;
private final DataSourceInternalWriterHelper dataSourceInternalWriterHelper;
public HoodieDataSourceInternalBatchWrite(String instantTime, HoodieWriteConfig writeConfig, StructType structType,
SparkSession jss, Configuration hadoopConfiguration) {
SparkSession jss, Configuration hadoopConfiguration, boolean arePartitionRecordsSorted) {
this.instantTime = instantTime;
this.writeConfig = writeConfig;
this.structType = structType;
this.arePartitionRecordsSorted = arePartitionRecordsSorted;
this.dataSourceInternalWriterHelper = new DataSourceInternalWriterHelper(instantTime, writeConfig, structType,
jss, hadoopConfiguration);
}
@@ -61,7 +63,7 @@ public class HoodieDataSourceInternalBatchWrite implements BatchWrite {
dataSourceInternalWriterHelper.createInflightCommit();
if (WriteOperationType.BULK_INSERT == dataSourceInternalWriterHelper.getWriteOperationType()) {
return new HoodieBulkInsertDataInternalWriterFactory(dataSourceInternalWriterHelper.getHoodieTable(),
writeConfig, instantTime, structType);
writeConfig, instantTime, structType, arePartitionRecordsSorted);
} else {
throw new IllegalArgumentException("Write Operation Type + " + dataSourceInternalWriterHelper.getWriteOperationType() + " not supported ");
}

View File

@@ -37,19 +37,21 @@ public class HoodieDataSourceInternalBatchWriteBuilder implements WriteBuilder {
private final StructType structType;
private final SparkSession jss;
private final Configuration hadoopConfiguration;
private final boolean arePartitionRecordsSorted;
public HoodieDataSourceInternalBatchWriteBuilder(String instantTime, HoodieWriteConfig writeConfig, StructType structType,
SparkSession jss, Configuration hadoopConfiguration) {
SparkSession jss, Configuration hadoopConfiguration, boolean arePartitionRecordsSorted) {
this.instantTime = instantTime;
this.writeConfig = writeConfig;
this.structType = structType;
this.jss = jss;
this.hadoopConfiguration = hadoopConfiguration;
this.arePartitionRecordsSorted = arePartitionRecordsSorted;
}
@Override
public BatchWrite buildForBatch() {
return new HoodieDataSourceInternalBatchWrite(instantTime, writeConfig, structType, jss,
hadoopConfiguration);
hadoopConfiguration, arePartitionRecordsSorted);
}
}

View File

@@ -41,14 +41,16 @@ class HoodieDataSourceInternalTable implements SupportsWrite {
private final StructType structType;
private final SparkSession jss;
private final Configuration hadoopConfiguration;
private final boolean arePartitionRecordsSorted;
public HoodieDataSourceInternalTable(String instantTime, HoodieWriteConfig config,
StructType schema, SparkSession jss, Configuration hadoopConfiguration) {
StructType schema, SparkSession jss, Configuration hadoopConfiguration, boolean arePartitionRecordsSorted) {
this.instantTime = instantTime;
this.writeConfig = config;
this.structType = schema;
this.jss = jss;
this.hadoopConfiguration = hadoopConfiguration;
this.arePartitionRecordsSorted = arePartitionRecordsSorted;
}
@Override
@@ -73,6 +75,6 @@ class HoodieDataSourceInternalTable implements SupportsWrite {
@Override
public WriteBuilder newWriteBuilder(LogicalWriteInfo logicalWriteInfo) {
return new HoodieDataSourceInternalBatchWriteBuilder(instantTime, writeConfig, structType, jss,
hadoopConfiguration);
hadoopConfiguration, arePartitionRecordsSorted);
}
}