lanyuanxiaoyao/hudi
1
0
Fork 0
Code Activity

[HUDI-2161] Adding support to disable meta columns with bulk insert operation (#3247)

Browse Source
This commit is contained in:
Sivabalan Narayanan
2021-07-19 20:43:48 -04:00
committed by GitHub
parent 2099bf41db
commit d5026e9a24
53 changed files with 1063 additions and 269 deletions
Download Patch File Download Diff File Expand all files Collapse all files

5
hudi-spark-datasource/hudi-spark3/src/main/java/org/apache/hudi/spark3/internal/DefaultSource.java
View File

@@ -19,6 +19,7 @@
package org.apache.hudi.spark3.internal;
import org.apache.hudi.DataSourceUtils;
import org.apache.hudi.common.table.HoodieTableConfig;
import org.apache.hudi.config.HoodieInternalConfig;
import org.apache.hudi.config.HoodieWriteConfig;
import org.apache.hudi.internal.BaseDefaultSource;
@@ -48,11 +49,13 @@ 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 populateMetaFields = Boolean.parseBoolean(properties.getOrDefault(HoodieTableConfig.HOODIE_POPULATE_META_FIELDS.key(),
HoodieTableConfig.HOODIE_POPULATE_META_FIELDS.defaultValue()));
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(), properties, arePartitionRecordsSorted);
getConfiguration(), properties, populateMetaFields, arePartitionRecordsSorted);
}
}

7
hudi-spark-datasource/hudi-spark3/src/main/java/org/apache/hudi/spark3/internal/HoodieBulkInsertDataInternalWriter.java
View File

@@ -37,9 +37,10 @@ public class HoodieBulkInsertDataInternalWriter implements DataWriter<InternalRo
private final BulkInsertDataInternalWriterHelper bulkInsertWriterHelper;
public HoodieBulkInsertDataInternalWriter(HoodieTable hoodieTable, HoodieWriteConfig writeConfig,
String instantTime, int taskPartitionId, long taskId, StructType structType, boolean arePartitionRecordsSorted) {
String instantTime, int taskPartitionId, long taskId, StructType structType, boolean populateMetaFields,
boolean arePartitionRecordsSorted) {
this.bulkInsertWriterHelper = new BulkInsertDataInternalWriterHelper(hoodieTable,
writeConfig, instantTime, taskPartitionId, taskId, 0, structType, arePartitionRecordsSorted);
writeConfig, instantTime, taskPartitionId, taskId, 0, structType, populateMetaFields, arePartitionRecordsSorted);
}
@Override
@@ -61,4 +62,4 @@ public class HoodieBulkInsertDataInternalWriter implements DataWriter<InternalRo
public void close() throws IOException {
bulkInsertWriterHelper.close();
}
}
}

7
hudi-spark-datasource/hudi-spark3/src/main/java/org/apache/hudi/spark3/internal/HoodieBulkInsertDataInternalWriterFactory.java
View File

@@ -35,20 +35,23 @@ public class HoodieBulkInsertDataInternalWriterFactory implements DataWriterFact
private final HoodieTable hoodieTable;
private final HoodieWriteConfig writeConfig;
private final StructType structType;
private final boolean populateMetaFields;
private final boolean arePartitionRecordsSorted;
public HoodieBulkInsertDataInternalWriterFactory(HoodieTable hoodieTable, HoodieWriteConfig writeConfig,
String instantTime, StructType structType, boolean arePartitionRecordsSorted) {
String instantTime, StructType structType, boolean populateMetaFields,
boolean arePartitionRecordsSorted) {
this.hoodieTable = hoodieTable;
this.writeConfig = writeConfig;
this.instantTime = instantTime;
this.structType = structType;
this.populateMetaFields = populateMetaFields;
this.arePartitionRecordsSorted = arePartitionRecordsSorted;
}
@Override
public DataWriter<InternalRow> createWriter(int partitionId, long taskId) {
return new HoodieBulkInsertDataInternalWriter(hoodieTable, writeConfig, instantTime, partitionId, taskId,
structType, arePartitionRecordsSorted);
structType, populateMetaFields, arePartitionRecordsSorted);
}
}

10
hudi-spark-datasource/hudi-spark3/src/main/java/org/apache/hudi/spark3/internal/HoodieDataSourceInternalBatchWrite.java
View File

@@ -27,10 +27,10 @@ import org.apache.hudi.internal.DataSourceInternalWriterHelper;
import org.apache.hadoop.conf.Configuration;
import org.apache.spark.sql.SparkSession;
import org.apache.spark.sql.connector.write.DataWriterFactory;
import org.apache.spark.sql.connector.write.WriterCommitMessage;
import org.apache.spark.sql.connector.write.BatchWrite;
import org.apache.spark.sql.connector.write.DataWriterFactory;
import org.apache.spark.sql.connector.write.PhysicalWriteInfo;
import org.apache.spark.sql.connector.write.WriterCommitMessage;
import org.apache.spark.sql.types.StructType;
import java.util.Arrays;
@@ -49,14 +49,16 @@ public class HoodieDataSourceInternalBatchWrite implements BatchWrite {
private final HoodieWriteConfig writeConfig;
private final StructType structType;
private final boolean arePartitionRecordsSorted;
private final boolean populateMetaFields;
private final DataSourceInternalWriterHelper dataSourceInternalWriterHelper;
private Map<String, String> extraMetadata = new HashMap<>();
public HoodieDataSourceInternalBatchWrite(String instantTime, HoodieWriteConfig writeConfig, StructType structType,
SparkSession jss, Configuration hadoopConfiguration, Map<String, String> properties, boolean arePartitionRecordsSorted) {
SparkSession jss, Configuration hadoopConfiguration, Map<String, String> properties, boolean populateMetaFields, boolean arePartitionRecordsSorted) {
this.instantTime = instantTime;
this.writeConfig = writeConfig;
this.structType = structType;
this.populateMetaFields = populateMetaFields;
this.arePartitionRecordsSorted = arePartitionRecordsSorted;
this.extraMetadata = DataSourceUtils.getExtraMetadata(properties);
this.dataSourceInternalWriterHelper = new DataSourceInternalWriterHelper(instantTime, writeConfig, structType,
@@ -68,7 +70,7 @@ public class HoodieDataSourceInternalBatchWrite implements BatchWrite {
dataSourceInternalWriterHelper.createInflightCommit();
if (WriteOperationType.BULK_INSERT == dataSourceInternalWriterHelper.getWriteOperationType()) {
return new HoodieBulkInsertDataInternalWriterFactory(dataSourceInternalWriterHelper.getHoodieTable(),
writeConfig, instantTime, structType, arePartitionRecordsSorted);
writeConfig, instantTime, structType, populateMetaFields, arePartitionRecordsSorted);
} else {
throw new IllegalArgumentException("Write Operation Type + " + dataSourceInternalWriterHelper.getWriteOperationType() + " not supported ");
}

11
hudi-spark-datasource/hudi-spark3/src/main/java/org/apache/hudi/spark3/internal/HoodieDataSourceInternalBatchWriteBuilder.java
View File

@@ -39,23 +39,26 @@ public class HoodieDataSourceInternalBatchWriteBuilder implements WriteBuilder {
private final StructType structType;
private final SparkSession jss;
private final Configuration hadoopConfiguration;
private final boolean arePartitionRecordsSorted;
private final Map<String, String> properties;
private final boolean populateMetaFields;
private final boolean arePartitionRecordsSorted;
public HoodieDataSourceInternalBatchWriteBuilder(String instantTime, HoodieWriteConfig writeConfig, StructType structType,
SparkSession jss, Configuration hadoopConfiguration, Map<String, String> properties, boolean arePartitionRecordsSorted) {
SparkSession jss, Configuration hadoopConfiguration, Map<String, String> properties, boolean populateMetaFields,
boolean arePartitionRecordsSorted) {
this.instantTime = instantTime;
this.writeConfig = writeConfig;
this.structType = structType;
this.jss = jss;
this.hadoopConfiguration = hadoopConfiguration;
this.arePartitionRecordsSorted = arePartitionRecordsSorted;
this.properties = properties;
this.populateMetaFields = populateMetaFields;
this.arePartitionRecordsSorted = arePartitionRecordsSorted;
}
@Override
public BatchWrite buildForBatch() {
return new HoodieDataSourceInternalBatchWrite(instantTime, writeConfig, structType, jss,
hadoopConfiguration, properties, arePartitionRecordsSorted);
hadoopConfiguration, properties, populateMetaFields, arePartitionRecordsSorted);
}
}

13
hudi-spark-datasource/hudi-spark3/src/main/java/org/apache/hudi/spark3/internal/HoodieDataSourceInternalTable.java
View File

@@ -18,9 +18,9 @@
package org.apache.hudi.spark3.internal;
import org.apache.hadoop.conf.Configuration;
import org.apache.hudi.config.HoodieWriteConfig;
import org.apache.hadoop.conf.Configuration;
import org.apache.spark.sql.SparkSession;
import org.apache.spark.sql.connector.catalog.SupportsWrite;
import org.apache.spark.sql.connector.catalog.TableCapability;
@@ -44,17 +44,19 @@ class HoodieDataSourceInternalTable implements SupportsWrite {
private final Configuration hadoopConfiguration;
private final boolean arePartitionRecordsSorted;
private final Map<String, String> properties;
private final boolean populateMetaFields;
public HoodieDataSourceInternalTable(String instantTime, HoodieWriteConfig config,
StructType schema, SparkSession jss, Configuration hadoopConfiguration, Map<String, String> properties,
boolean arePartitionRecordsSorted) {
boolean populateMetaFields, boolean arePartitionRecordsSorted) {
this.instantTime = instantTime;
this.writeConfig = config;
this.structType = schema;
this.jss = jss;
this.hadoopConfiguration = hadoopConfiguration;
this.arePartitionRecordsSorted = arePartitionRecordsSorted;
this.properties = properties;
this.populateMetaFields = populateMetaFields;
this.arePartitionRecordsSorted = arePartitionRecordsSorted;
}
@Override
@@ -69,7 +71,8 @@ class HoodieDataSourceInternalTable implements SupportsWrite {
@Override
public Set<TableCapability> capabilities() {
return new HashSet<TableCapability>() {{
return new HashSet<TableCapability>() {
{
add(TableCapability.BATCH_WRITE);
add(TableCapability.TRUNCATE);
}
@@ -79,6 +82,6 @@ class HoodieDataSourceInternalTable implements SupportsWrite {
@Override
public WriteBuilder newWriteBuilder(LogicalWriteInfo logicalWriteInfo) {
return new HoodieDataSourceInternalBatchWriteBuilder(instantTime, writeConfig, structType, jss,
hadoopConfiguration, properties, arePartitionRecordsSorted);
hadoopConfiguration, properties, populateMetaFields, arePartitionRecordsSorted);
}
}