[GitHub] [iceberg] stevenzwu commented on a change in pull request #2305: Flink: FLIP-27 source split and reader

[GitBox <gi...@apache.org>]

stevenzwu commented on a change in pull request #2305:
URL: https://github.com/apache/iceberg/pull/2305#discussion_r593986709



##########
File path: flink/src/main/java/org/apache/iceberg/flink/source/reader/FlinkBulkFormatAdaptor.java
##########
@@ -0,0 +1,222 @@
+/*
+ * 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.source.reader;
+
+import java.io.IOException;
+import java.net.URI;
+import java.util.Iterator;
+import java.util.List;
+import java.util.Map;
+import java.util.stream.Collectors;
+import javax.annotation.Nullable;
+import org.apache.flink.api.common.typeinfo.TypeInformation;
+import org.apache.flink.configuration.Configuration;
+import org.apache.flink.connector.file.src.FileSourceSplit;
+import org.apache.flink.connector.file.src.reader.BulkFormat;
+import org.apache.flink.connector.file.src.util.CheckpointedPosition;
+import org.apache.flink.connector.file.src.util.MutableRecordAndPosition;
+import org.apache.flink.connector.file.src.util.RecordAndPosition;
+import org.apache.flink.core.fs.Path;
+import org.apache.iceberg.FileFormat;
+import org.apache.iceberg.FileScanTask;
+import org.apache.iceberg.flink.source.split.IcebergSourceSplit;
+import org.apache.iceberg.relocated.com.google.common.base.Preconditions;
+
+/**
+ * This adaptor use Flink BulkFormat implementation to read data file.
+ * Note that Flink BulkFormat may not support row deletes like {@link RowDataIteratorBulkFormat}
+ */
+public class FlinkBulkFormatAdaptor<T> implements BulkFormat<T, IcebergSourceSplit> {
+
+  private final Map<FileFormat, BulkFormat<T, FileSourceSplit>> bulkFormatProvider;
+  private final TypeInformation<T> producedType;
+
+  public FlinkBulkFormatAdaptor(Map<FileFormat, BulkFormat<T, FileSourceSplit>> bulkFormatProvider) {
+    this.bulkFormatProvider = bulkFormatProvider;
+    // validate that all BulkFormat produce the same type
+    List<TypeInformation<T>> uniqueTypes = bulkFormatProvider.values().stream()
+        .map(bulkFormat -> bulkFormat.getProducedType())
+        .distinct()
+        .collect(Collectors.toList());
+    Preconditions.checkArgument(uniqueTypes.size() == 1,
+        "BulkFormats have the different producedType: " + uniqueTypes);
+    producedType = uniqueTypes.get(0);
+  }
+
+  @Override
+  public Reader<T> createReader(Configuration config, IcebergSourceSplit split) throws IOException {
+    return new ReaderAdaptor<T>(bulkFormatProvider, config, split, false);
+  }
+
+  @Override
+  public Reader<T> restoreReader(Configuration config, IcebergSourceSplit split) throws IOException {
+    return new ReaderAdaptor<T>(bulkFormatProvider, config, split, true);
+  }
+
+  @Override
+  public boolean isSplittable() {
+    return false;
+  }
+
+  @Override
+  public TypeInformation<T> getProducedType() {
+    return producedType;
+  }
+
+  private static final class ReaderAdaptor<T> implements BulkFormat.Reader<T> {
+
+    private final Map<FileFormat, BulkFormat<T, FileSourceSplit>> bulkFormatProvider;
+    private final Configuration config;
+    private final Iterator<FileScanTask> fileIterator;
+    private final boolean isRestored;
+
+    // file offset in CombinedScanTask
+    private int fileOffset = -1;
+    private Reader<T> currentReader;
+
+    ReaderAdaptor(
+        Map<FileFormat, BulkFormat<T, FileSourceSplit>> bulkFormatProvider,
+        Configuration config,
+        IcebergSourceSplit icebergSplit,
+        boolean isRestored) throws IOException {
+      this.config = config;
+      this.bulkFormatProvider = bulkFormatProvider;
+      this.fileIterator = icebergSplit.task().files().iterator();
+      this.isRestored = isRestored;
+
+      final CheckpointedPosition position = icebergSplit.checkpointedPosition();
+      if (position != null) {
+        // skip files based on offset in checkpointed position
+        Preconditions.checkArgument(position.getOffset() < icebergSplit.task().files().size(),
+            String.format("Checkpointed file offset is %d, while CombinedScanTask has %d files",
+                position.getOffset(), icebergSplit.task().files().size()));
+        for (int i = 0; i < position.getOffset(); ++i) {
+          fileIterator.next();
+          fileOffset++;
+        }
+        // first file may need to skip records
+        setupReader(position.getRecordsAfterOffset());
+      } else {
+        setupReader(0L);
+      }
+    }
+
+    /**
+     * TODO: we can't return RecordIterator with empty data.
+     * Otherwise, caller may assume it is end of input.
+     * We probably need to add a {@code hasNext()} API to
+     * {@link RecordIterator} to achieve the goal.
+     */
+    @Nullable
+    @Override
+    public RecordIterator<T> readBatch() throws IOException {
+      RecordIterator<T> result = null;
+      while (currentReader != null || fileIterator.hasNext()) {
+        if (currentReader == null) {
+          setupReader(0L);
+        } else {
+          result = currentReader.readBatch();
+          if (result != null) {
+            break;
+          } else {
+            closeCurrentReader();
+          }
+        }
+      }
+      if (result == null) {
+        return null;
+      } else {
+        return new RecordIteratorAdaptor(fileOffset, result);
+      }
+    }
+
+    @Override
+    public void close() throws IOException {
+      closeCurrentReader();
+    }
+
+    private void closeCurrentReader() throws IOException {
+      if (currentReader != null) {
+        currentReader.close();
+        currentReader = null;
+      }
+    }
+
+    private void setupReader(long skipRecordCount) throws IOException {
+      if (fileIterator.hasNext()) {
+        final FileScanTask fileScanTask = fileIterator.next();
+        final FileFormat fileFormat = fileScanTask.file().format();
+        if (!bulkFormatProvider.containsKey(fileFormat)) {
+          throw new IOException("Unsupported file format: " + fileFormat);
+        }
+        final BulkFormat<T, FileSourceSplit> bulkFormat = bulkFormatProvider.get(fileFormat);
+        fileOffset++;
+        final FileSourceSplit fileSourceSplit = new FileSourceSplit(
+            "",
+            new Path(URI.create(fileScanTask.file().path().toString())),

Review comment:
       We extended from `FileSourceSplit` mainly for the `BulkFormat` batch reader interface so that we can plug in vectorized readers from Flink. I am also debating if this is the right thing to do as mentioned in the original description.
   
   But this is not really relevant to FileIO, which deals with underneath filesystem like S3. Here we are mainly talking about file format reader (like Parquet, Orc).




----------------------------------------------------------------
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:
users@infra.apache.org



---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@iceberg.apache.org
For additional commands, e-mail: issues-help@iceberg.apache.org