[GitHub] [iceberg] rdblue commented on a change in pull request #1254: Parquet: Add row position reader

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

rdblue commented on a change in pull request #1254:
URL: https://github.com/apache/iceberg/pull/1254#discussion_r461890941



##########
File path: parquet/src/main/java/org/apache/iceberg/parquet/ParquetValueReader.java
##########
@@ -30,4 +30,6 @@
   List<TripleIterator<?>> columns();
 
   void setPageSource(PageReadStore pageStore);
+
+  default void setRowOffsetForRowGroup(long position) {}

Review comment:
       Why not add the position to the page source? Then the two operations are tied together: the row offset is the start offset for the new pages.

##########
File path: parquet/src/main/java/org/apache/iceberg/parquet/ParquetValueReaders.java
##########
@@ -137,6 +137,41 @@ public void setPageSource(PageReadStore pageStore) {
     }
   }
 
+  static class PositionReader implements ParquetValueReader<Long> {
+    private long rowOffsetInCurrentRowGroup = -1;
+    private long rowGroupRowOffsetInFile;
+
+    @Override
+    public Long read(Long reuse) {
+      rowOffsetInCurrentRowGroup = rowOffsetInCurrentRowGroup + 1;
+      return rowGroupRowOffsetInFile + rowOffsetInCurrentRowGroup;
+    }
+
+    @Override
+    public TripleIterator<?> column() {
+      return NullReader.NULL_COLUMN;
+    }
+
+    @Override
+    public List<TripleIterator<?>> columns() {
+      return NullReader.COLUMNS;
+    }
+
+    @Override
+    public void setPageSource(PageReadStore pageStore) {
+    }
+
+    @Override
+    public void setRowOffsetForRowGroup(long rowGroupStartPos) {
+      this.rowGroupRowOffsetInFile = rowGroupStartPos;
+      this.rowOffsetInCurrentRowGroup = -1;
+    }
+  }
+
+  public static ParquetValueReader<Long> position() {
+    return new PositionReader();
+  }

Review comment:
       Can you move this to the top of the file with the other factory methods?

##########
File path: parquet/src/main/java/org/apache/iceberg/parquet/ReadConf.java
##########
@@ -157,6 +164,23 @@ ParquetFileReader reader() {
     return shouldSkip;
   }
 
+  private Map<Long, Long> generateRowGroupsStartRowPos() {
+    ParquetFileReader fileReader = newReader(this.file, ParquetReadOptions.builder().build());
+    Map<Long, Long> offsetToStartRowPosMap = new HashMap<>();
+    long curRowCount = 0;
+    for (int i = 0; i < fileReader.getRowGroups().size(); i += 1) {
+      BlockMetaData meta = fileReader.getRowGroups().get(i);
+      offsetToStartRowPosMap.put(meta.getStartingPos(), curRowCount);
+      curRowCount += meta.getRowCount();
+    }
+
+    return offsetToStartRowPosMap;
+  }
+
+  long[] getRowGroupsStartRowPos() {

Review comment:
       How about naming this `startPositions`?

##########
File path: parquet/src/main/java/org/apache/iceberg/parquet/ReadConf.java
##########
@@ -157,6 +164,23 @@ ParquetFileReader reader() {
     return shouldSkip;
   }
 
+  private Map<Long, Long> generateRowGroupsStartRowPos() {

Review comment:
       Why does this separately read the Parquet file to create a map that is used to initialize an array, when the starting position could be set for the array in the existing loop? I don't think this method is needed.

##########
File path: parquet/src/main/java/org/apache/iceberg/parquet/ParquetValueReaders.java
##########
@@ -137,6 +137,41 @@ public void setPageSource(PageReadStore pageStore) {
     }
   }
 
+  static class PositionReader implements ParquetValueReader<Long> {
+    private long rowOffsetInCurrentRowGroup = -1;
+    private long rowGroupRowOffsetInFile;

Review comment:
       In general, try to be specific with names, but avoid unnecessary context. In this case, these names can be simpler: `rowGroupStart` and `rowOffset` would work fine. Extra context like `InFile` and `InCurrent` aren't adding clarity.

##########
File path: spark/src/test/java/org/apache/iceberg/spark/data/TestSparkParquetReadMetadataColumns.java
##########
@@ -0,0 +1,227 @@
+/*
+ * 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.spark.data;
+
+import java.io.File;
+import java.io.IOException;
+import java.util.ArrayList;
+import java.util.Iterator;
+import java.util.List;
+import org.apache.arrow.vector.NullCheckingForGet;
+import org.apache.hadoop.conf.Configuration;
+import org.apache.hadoop.fs.Path;
+import org.apache.iceberg.Files;
+import org.apache.iceberg.MetadataColumns;
+import org.apache.iceberg.Schema;
+import org.apache.iceberg.expressions.Expression;
+import org.apache.iceberg.expressions.Expressions;
+import org.apache.iceberg.io.CloseableIterable;
+import org.apache.iceberg.io.FileAppender;
+import org.apache.iceberg.parquet.Parquet;
+import org.apache.iceberg.parquet.ParquetSchemaUtil;
+import org.apache.iceberg.relocated.com.google.common.collect.Lists;
+import org.apache.iceberg.spark.SparkSchemaUtil;
+import org.apache.iceberg.spark.data.vectorized.VectorizedSparkParquetReaders;
+import org.apache.iceberg.types.Types;
+import org.apache.parquet.ParquetReadOptions;
+import org.apache.parquet.hadoop.ParquetFileReader;
+import org.apache.parquet.hadoop.ParquetFileWriter;
+import org.apache.parquet.hadoop.metadata.BlockMetaData;
+import org.apache.parquet.hadoop.util.HadoopInputFile;
+import org.apache.spark.sql.catalyst.InternalRow;
+import org.apache.spark.sql.catalyst.expressions.GenericInternalRow;
+import org.apache.spark.sql.types.StructType;
+import org.apache.spark.unsafe.types.UTF8String;
+import org.junit.Assert;
+import org.junit.Before;
+import org.junit.Rule;
+import org.junit.Test;
+import org.junit.rules.TemporaryFolder;
+import org.junit.runner.RunWith;
+import org.junit.runners.Parameterized;
+
+import static org.apache.iceberg.types.Types.NestedField.required;
+
+@RunWith(Parameterized.class)
+public class TestSparkParquetReadMetadataColumns {
+  private static final Schema DATA_SCHEMA = new Schema(
+      required(100, "id", Types.LongType.get()),
+      required(101, "data", Types.StringType.get())
+  );
+
+  private static final Schema PROJECTION_SCHEMA = new Schema(
+      required(100, "id", Types.LongType.get()),
+      required(101, "data", Types.StringType.get()),
+      MetadataColumns.ROW_POSITION
+  );
+
+  private static final int NUM_ROWS = 1000;
+  private static final List<InternalRow> DATA_ROWS;
+  private static final List<InternalRow> EXPECTED_ROWS;
+  private static final int NUM_ROW_GROUPS = 10;
+  private static final int ROWS_PER_SPLIT = NUM_ROWS / NUM_ROW_GROUPS;
+
+  static {
+    DATA_ROWS = Lists.newArrayListWithCapacity(NUM_ROWS);
+    for (long i = 0; i < NUM_ROWS; i += 1) {
+      InternalRow row = new GenericInternalRow(DATA_SCHEMA.columns().size());
+      if (i >= 500) {
+        row.update(0, 2 * i);
+      } else {
+        row.update(0, i);
+      }
+      row.update(1, UTF8String.fromString("str" + i));
+      DATA_ROWS.add(row);
+    }
+
+    EXPECTED_ROWS = Lists.newArrayListWithCapacity(NUM_ROWS);
+    for (long i = 0; i < NUM_ROWS; i += 1) {
+      InternalRow row = new GenericInternalRow(PROJECTION_SCHEMA.columns().size());
+      if (i >= 500) {
+        row.update(0, 2 * i);
+      } else {
+        row.update(0, i);
+      }
+      row.update(1, UTF8String.fromString("str" + i));
+      row.update(2, i);
+      EXPECTED_ROWS.add(row);
+    }
+  }
+
+  @Parameterized.Parameters
+  public static Object[][] parameters() {
+    return new Object[][] {
+        new Object[] { false },
+        //new Object[] { true }
+    };
+  }
+
+  @Rule
+  public TemporaryFolder temp = new TemporaryFolder();
+
+  private boolean vectorized;
+  private File testFile;
+
+  public TestSparkParquetReadMetadataColumns(boolean vectorized) {
+    this.vectorized = vectorized;
+  }
+
+  @Before
+  public void writeFile() throws IOException {
+    List<Path> fileSplits = new ArrayList<>();
+    StructType struct = SparkSchemaUtil.convert(DATA_SCHEMA);
+    Configuration conf = new Configuration();
+
+    testFile = temp.newFile();
+    Assert.assertTrue("Delete should succeed", testFile.delete());
+    ParquetFileWriter parquetFileWriter = new ParquetFileWriter(
+        conf,
+        ParquetSchemaUtil.convert(DATA_SCHEMA, "testSchema"),
+        new Path(testFile.getAbsolutePath())
+    );
+
+    parquetFileWriter.start();
+    for (int i = 0; i < NUM_ROW_GROUPS; i += 1) {
+      File split = temp.newFile();
+      Assert.assertTrue("Delete should succeed", split.delete());
+      fileSplits.add(new Path(split.getAbsolutePath()));
+      try (FileAppender<InternalRow> writer = Parquet.write(Files.localOutput(split))
+          .createWriterFunc(msgType -> SparkParquetWriters.buildWriter(struct, msgType))
+          .schema(DATA_SCHEMA)
+          .overwrite()
+          .build()) {
+        writer.addAll(DATA_ROWS.subList(i * ROWS_PER_SPLIT, (i + 1) * ROWS_PER_SPLIT));
+      }
+      parquetFileWriter.appendFile(HadoopInputFile.fromPath(new Path(split.getAbsolutePath()), conf));
+    }
+    parquetFileWriter
+        .end(ParquetFileWriter.mergeMetadataFiles(fileSplits, conf).getFileMetaData().getKeyValueMetaData());
+  }
+
+  @Test
+  public void testReadRowNumbers() throws IOException {
+    readAndValidate(null, null, null, EXPECTED_ROWS);
+  }
+
+  @Test
+  public void testReadRowNumbersWithFilter() throws IOException {
+    // current iceberg supports row group filter.
+    for (int i = 1; i < 5; i += 1) {
+      readAndValidate(
+          Expressions.and(Expressions.lessThan("id", 500),
+              Expressions.greaterThanOrEqual("id", i * ROWS_PER_SPLIT)),
+          null,
+          null,
+          EXPECTED_ROWS.subList(i * ROWS_PER_SPLIT, 500));
+    }
+  }
+
+  @Test
+  public void testReadRowNumbersWithSplits() throws IOException {
+    ParquetFileReader fileReader = new ParquetFileReader(
+        HadoopInputFile.fromPath(new Path(testFile.getAbsolutePath()), new Configuration()),
+        ParquetReadOptions.builder().build());
+    List<BlockMetaData> rowGroups = fileReader.getRowGroups();
+    for (int i = 0; i < NUM_ROW_GROUPS; i += 1) {
+      readAndValidate(null,
+          rowGroups.get(i).getColumns().get(0).getStartingPos(),
+          rowGroups.get(i).getCompressedSize(),
+          EXPECTED_ROWS.subList(i * ROWS_PER_SPLIT, (i + 1) * ROWS_PER_SPLIT));
+    }
+  }
+
+  private void readAndValidate(Expression filter, Long splitStart, Long splitLength, List<InternalRow> expected)
+      throws IOException {
+    CloseableIterable<InternalRow> reader = null;
+    try {
+      Parquet.ReadBuilder builder = Parquet.read(Files.localInput(testFile))
+          .project(PROJECTION_SCHEMA);
+
+      if (vectorized) {
+        builder.createBatchedReaderFunc(fileSchema -> VectorizedSparkParquetReaders.buildReader(PROJECTION_SCHEMA,
+            fileSchema, NullCheckingForGet.NULL_CHECKING_ENABLED));
+      } else {
+        builder = builder.createReaderFunc(msgType -> SparkParquetReaders.buildReader(PROJECTION_SCHEMA, msgType));
+      }
+
+      if (filter != null) {
+        builder = builder.filter(filter);
+      }
+
+      if (splitStart != null && splitLength != null) {
+        builder = builder.split(splitStart, splitLength);
+      }
+
+      reader = builder.build();
+
+      final Iterator<InternalRow> actualRows = reader.iterator();
+
+      for (InternalRow internalRow : expected) {
+        Assert.assertTrue("Should have expected number of rows", actualRows.hasNext());
+        TestHelpers.assertEquals(PROJECTION_SCHEMA, internalRow, actualRows.next());
+      }
+      Assert.assertFalse("Should not have extra rows", actualRows.hasNext());
+    } finally {
+      if (reader != null) {
+        reader.close();

Review comment:
       Why not use try-with-resources instead of a `finally` block?




----------------------------------------------------------------
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