Skip to content
Merged
Show file tree
Hide file tree
Changes from 1 commit
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
Original file line number Diff line number Diff line change
Expand Up @@ -57,7 +57,8 @@ private FormatModelRegistry() {}
private static final List<String> CLASSES_TO_REGISTER =
ImmutableList.of(
"org.apache.iceberg.data.GenericFormatModels",
"org.apache.iceberg.arrow.vectorized.ArrowFormatModels");
"org.apache.iceberg.arrow.vectorized.ArrowFormatModels",
"org.apache.iceberg.flink.data.FlinkFormatModels");

// Format models indexed by file format and object model class
private static final Map<Pair<FileFormat, Class<?>>, FormatModel<?, ?>> MODELS =
Expand Down
Original file line number Diff line number Diff line change
@@ -0,0 +1,182 @@
/*
* Licensed to the Apache Software Foundation (ASF) under one
* or more contributor license agreements. See the NOTICE file
* distributed with this work for additional information
* regarding copyright ownership. The ASF licenses this file
* to you under the Apache License, Version 2.0 (the
* "License"); you may not use this file except in compliance
* with the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing,
* software distributed under the License is distributed on an
* "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
* KIND, either express or implied. See the License for the
* specific language governing permissions and limitations
* under the License.
*/
package org.apache.iceberg.data;

import java.io.IOException;
import java.io.Serializable;
import java.io.UncheckedIOException;
import java.util.Map;
import org.apache.iceberg.FileFormat;
import org.apache.iceberg.MetricsConfig;
import org.apache.iceberg.PartitionSpec;
import org.apache.iceberg.Schema;
import org.apache.iceberg.SortOrder;
import org.apache.iceberg.StructLike;
import org.apache.iceberg.Table;
import org.apache.iceberg.deletes.EqualityDeleteWriter;
import org.apache.iceberg.deletes.PositionDeleteWriter;
import org.apache.iceberg.encryption.EncryptedOutputFile;
import org.apache.iceberg.encryption.EncryptionKeyMetadata;
import org.apache.iceberg.formats.FileWriterBuilder;
import org.apache.iceberg.formats.FormatModelRegistry;
import org.apache.iceberg.io.DataWriter;
import org.apache.iceberg.io.FileWriterFactory;
import org.apache.iceberg.relocated.com.google.common.base.Preconditions;
import org.apache.iceberg.relocated.com.google.common.collect.ImmutableMap;

/**
* A base writer factory to be extended by query engine integrations.
*
* @param <T> row type
*/
public abstract class RegistryBasedFileWriterFactory<T, S>
implements FileWriterFactory<T>, Serializable {
private final Table table;
private final FileFormat dataFileFormat;
private final Class<T> inputType;
private final Schema dataSchema;
private final SortOrder dataSortOrder;
private final FileFormat deleteFileFormat;
private final int[] equalityFieldIds;
private final Schema equalityDeleteRowSchema;
private final SortOrder equalityDeleteSortOrder;
private final Map<String, String> writerProperties;
private final S inputSchema;
private final S equalityDeleteInputSchema;

protected RegistryBasedFileWriterFactory(
Table table,
FileFormat dataFileFormat,
Class<T> inputType,
Schema dataSchema,
SortOrder dataSortOrder,
FileFormat deleteFileFormat,
int[] equalityFieldIds,
Schema equalityDeleteRowSchema,
SortOrder equalityDeleteSortOrder,
Map<String, String> writerProperties,
S inputSchema,
S equalityDeleteInputSchema) {
this.table = table;
this.dataFileFormat = dataFileFormat;
this.inputType = inputType;
this.dataSchema = dataSchema;
this.dataSortOrder = dataSortOrder;
this.deleteFileFormat = deleteFileFormat;
this.equalityFieldIds = equalityFieldIds;
this.equalityDeleteRowSchema = equalityDeleteRowSchema;
this.equalityDeleteSortOrder = equalityDeleteSortOrder;
this.writerProperties = writerProperties != null ? writerProperties : ImmutableMap.of();
this.inputSchema = inputSchema;
this.equalityDeleteInputSchema = equalityDeleteInputSchema;
}

protected S inputSchema() {
return inputSchema;
}

protected S equalityDeleteInputSchema() {
return equalityDeleteInputSchema;
}

@Override
public DataWriter<T> newDataWriter(
EncryptedOutputFile file, PartitionSpec spec, StructLike partition) {
Preconditions.checkNotNull(dataSchema, "Data schema must not be null");
EncryptionKeyMetadata keyMetadata = file.keyMetadata();
Map<String, String> properties = table != null ? table.properties() : ImmutableMap.of();
MetricsConfig metricsConfig =
table != null ? MetricsConfig.forTable(table) : MetricsConfig.getDefault();

try {
FileWriterBuilder<DataWriter<T>, S> builder =
FormatModelRegistry.dataWriteBuilder(dataFileFormat, inputType, file);
return builder
.schema(dataSchema)
.engineSchema(inputSchema())
.setAll(properties)
.setAll(writerProperties)
.metricsConfig(metricsConfig)
.spec(spec)
.partition(partition)
.keyMetadata(keyMetadata)
.sortOrder(dataSortOrder)
.overwrite()
.build();
} catch (IOException e) {
throw new UncheckedIOException("Failed to create new data writer", e);
}
}

@Override
public EqualityDeleteWriter<T> newEqualityDeleteWriter(
EncryptedOutputFile file, PartitionSpec spec, StructLike partition) {
Preconditions.checkNotNull(equalityDeleteRowSchema, "Equality delete schema must not be null");

EncryptionKeyMetadata keyMetadata = file.keyMetadata();
Map<String, String> properties = table != null ? table.properties() : ImmutableMap.of();
MetricsConfig metricsConfig =
table != null ? MetricsConfig.forTable(table) : MetricsConfig.getDefault();

try {
FileWriterBuilder<EqualityDeleteWriter<T>, S> builder =
FormatModelRegistry.equalityDeleteWriteBuilder(deleteFileFormat, inputType, file);
return builder
.setAll(properties)
.setAll(writerProperties)
.metricsConfig(metricsConfig)
.schema(equalityDeleteRowSchema)
.engineSchema(equalityDeleteInputSchema())
.equalityFieldIds(equalityFieldIds)
.spec(spec)
.partition(partition)
.keyMetadata(keyMetadata)
.sortOrder(equalityDeleteSortOrder)
.overwrite()
.build();
} catch (IOException e) {
throw new UncheckedIOException("Failed to create new equality delete writer", e);
}
}

@Override
public PositionDeleteWriter<T> newPositionDeleteWriter(
EncryptedOutputFile file, PartitionSpec spec, StructLike partition) {
EncryptionKeyMetadata keyMetadata = file.keyMetadata();
Map<String, String> properties = table != null ? table.properties() : ImmutableMap.of();
MetricsConfig metricsConfig =
table != null ? MetricsConfig.forPositionDelete(table) : MetricsConfig.forPositionDelete();

try {
FileWriterBuilder<PositionDeleteWriter<T>, ?> builder =
FormatModelRegistry.positionDeleteWriteBuilder(deleteFileFormat, file);
return builder
.setAll(properties)
.setAll(writerProperties)
.metricsConfig(metricsConfig)
.spec(spec)
.partition(partition)
.keyMetadata(keyMetadata)
.overwrite()
.build();
} catch (IOException e) {
throw new UncheckedIOException("Failed to create new position delete writer", e);
}
}
}
Original file line number Diff line number Diff line change
@@ -0,0 +1,58 @@
/*
* Licensed to the Apache Software Foundation (ASF) under one
* or more contributor license agreements. See the NOTICE file
* distributed with this work for additional information
* regarding copyright ownership. The ASF licenses this file
* to you under the Apache License, Version 2.0 (the
* "License"); you may not use this file except in compliance
* with the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing,
* software distributed under the License is distributed on an
* "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
* KIND, either express or implied. See the License for the
* specific language governing permissions and limitations
* under the License.
*/
package org.apache.iceberg.flink.data;

import org.apache.flink.table.data.RowData;
import org.apache.flink.table.types.logical.RowType;
import org.apache.iceberg.avro.AvroFormatModel;
import org.apache.iceberg.formats.FormatModelRegistry;
import org.apache.iceberg.orc.ORCFormatModel;
import org.apache.iceberg.parquet.ParquetFormatModel;

public class FlinkFormatModels {
public static void register() {
FormatModelRegistry.register(
ParquetFormatModel.create(
RowData.class,
RowType.class,
(icebergSchema, fileSchema, engineSchema) ->
FlinkParquetWriters.buildWriter(engineSchema, fileSchema),
(icebergSchema, fileSchema, engineSchema, idToConstant) ->
FlinkParquetReaders.buildReader(icebergSchema, fileSchema, idToConstant)));

FormatModelRegistry.register(
AvroFormatModel.create(
RowData.class,
RowType.class,
(icebergSchema, fileSchema, engineSchema) -> new FlinkAvroWriter(engineSchema),
(icebergSchema, fileSchema, engineSchema, idToConstant) ->
FlinkPlannedAvroReader.create(icebergSchema, idToConstant)));

FormatModelRegistry.register(
ORCFormatModel.create(
RowData.class,
RowType.class,
(icebergSchema, fileSchema, engineSchema) ->
FlinkOrcWriter.buildWriter(engineSchema, icebergSchema),
(icebergSchema, fileSchema, engineSchema, idToConstant) ->
new FlinkOrcReader(icebergSchema, fileSchema, idToConstant)));
}

private FlinkFormatModels() {}
}
Original file line number Diff line number Diff line change
Expand Up @@ -19,6 +19,7 @@
package org.apache.iceberg.flink.data;

import java.util.List;
import org.apache.flink.annotation.Internal;
import org.apache.flink.table.types.logical.ArrayType;
import org.apache.flink.table.types.logical.LogicalType;
import org.apache.flink.table.types.logical.MapType;
Expand All @@ -29,9 +30,10 @@
import org.apache.iceberg.types.Type;
import org.apache.iceberg.types.Types;

@Internal
abstract class FlinkSchemaVisitor<T> {

static <T> T visit(RowType flinkType, Schema schema, FlinkSchemaVisitor<T> visitor) {
public static <T> T visit(RowType flinkType, Schema schema, FlinkSchemaVisitor<T> visitor) {
Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

is this public change required?

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Nope. Remained from old versions.
Reverted

return visit(flinkType, schema.asStruct(), visitor);
}

Expand Down Expand Up @@ -94,24 +96,29 @@ private static <T> T visitRecord(
List<LogicalType> fieldTypes = Lists.newArrayListWithExpectedSize(fieldSize);
List<Types.NestedField> nestedFields = struct.fields();

for (int i = 0; i < fieldSize; i++) {
Types.NestedField iField = nestedFields.get(i);
int fieldIndex = rowType.getFieldIndex(iField.name());
Preconditions.checkArgument(
fieldIndex >= 0, "NestedField: %s is not found in flink RowType: %s", iField, rowType);
visitor.beforeStruct(struct.asStructType());

LogicalType fieldFlinkType = rowType.getTypeAt(fieldIndex);
try {
Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

why do we need to switch the visit impl to try-finally? if there is any exception, it would just fail. is it important to call the after methods?

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

We have the same pattern for every field, that the after method is called in the finally. Since I have added the afterStruct, I have decided to follow the same pattern here as well.

for (int i = 0; i < fieldSize; i++) {
Types.NestedField iField = nestedFields.get(i);
int fieldIndex = rowType.getFieldIndex(iField.name());
Preconditions.checkArgument(
fieldIndex >= 0, "NestedField: %s is not found in flink RowType: %s", iField, rowType);

fieldTypes.add(fieldFlinkType);
LogicalType fieldFlinkType = rowType.getTypeAt(fieldIndex);
fieldTypes.add(fieldFlinkType);

visitor.beforeField(iField);
try {
if (iField.type() != Types.UnknownType.get()) {
results.add(visit(fieldFlinkType, iField.type(), visitor));
visitor.beforeField(iField);
try {
if (iField.type() != Types.UnknownType.get()) {
results.add(visit(fieldFlinkType, iField.type(), visitor));
}
} finally {
visitor.afterField(iField);
}
} finally {
visitor.afterField(iField);
}
} finally {
visitor.afterStruct(struct.asStructType());
}

return visitor.record(struct, results, fieldTypes);
Expand All @@ -137,6 +144,10 @@ public void beforeField(Types.NestedField field) {}

public void afterField(Types.NestedField field) {}

public void beforeStruct(Types.StructType type) {}

public void afterStruct(Types.StructType type) {}

public void beforeListElement(Types.NestedField elementField) {
beforeField(elementField);
}
Expand Down
Loading