[GitHub] [incubator-iceberg] samarthjain commented on a change in pull request #710: Parquet changes for vectorized reads

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

samarthjain commented on a change in pull request #710: Parquet changes for vectorized reads
URL: https://github.com/apache/incubator-iceberg/pull/710#discussion_r361555575
 
 

 ##########
 File path: parquet/src/main/java/org/apache/iceberg/parquet/vectorized/VectorizedParquetReader.java
 ##########
 @@ -0,0 +1,290 @@
+/*
+ * 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.parquet.vectorized;
+
+import java.io.Closeable;
+import java.io.IOException;
+import java.util.HashMap;
+import java.util.Iterator;
+import java.util.List;
+import java.util.Map;
+import java.util.NoSuchElementException;
+import java.util.function.Function;
+import org.apache.iceberg.Schema;
+import org.apache.iceberg.exceptions.RuntimeIOException;
+import org.apache.iceberg.expressions.Expression;
+import org.apache.iceberg.expressions.Expressions;
+import org.apache.iceberg.io.CloseableGroup;
+import org.apache.iceberg.io.CloseableIterable;
+import org.apache.iceberg.io.InputFile;
+import org.apache.iceberg.parquet.ParquetDictionaryRowGroupFilter;
+import org.apache.iceberg.parquet.ParquetIO;
+import org.apache.iceberg.parquet.ParquetMetricsRowGroupFilter;
+import org.apache.iceberg.parquet.ParquetSchemaUtil;
+import org.apache.iceberg.parquet.ParquetUtil;
+import org.apache.parquet.ParquetReadOptions;
+import org.apache.parquet.column.page.DictionaryPageReadStore;
+import org.apache.parquet.column.page.PageReadStore;
+import org.apache.parquet.hadoop.ParquetFileReader;
+import org.apache.parquet.hadoop.metadata.BlockMetaData;
+import org.apache.parquet.hadoop.metadata.ColumnChunkMetaData;
+import org.apache.parquet.hadoop.metadata.ColumnPath;
+import org.apache.parquet.schema.MessageType;
+
+public class VectorizedParquetReader<T> extends CloseableGroup implements CloseableIterable<T> {
+  private final InputFile input;
+  private final Schema expectedSchema;
+  private final ParquetReadOptions options;
+  private final Function<MessageType, VectorizedReader<T>> batchReaderFunc;
+  private final Expression filter;
+  private final boolean reuseContainers;
+  private final boolean caseSensitive;
+  private final int batchSize;
+
+  public VectorizedParquetReader(
+      InputFile input, Schema expectedSchema, ParquetReadOptions options,
+      Function<MessageType, VectorizedReader<T>> readerFunc,
+      Expression filter, boolean reuseContainers, boolean caseSensitive, int maxRecordsPerBatch) {
+    this.input = input;
+    this.expectedSchema = expectedSchema;
+    this.options = options;
+    this.batchReaderFunc = readerFunc;
+    // replace alwaysTrue with null to avoid extra work evaluating a trivial filter
+    this.filter = filter == Expressions.alwaysTrue() ? null : filter;
+    this.reuseContainers = reuseContainers;
+    this.caseSensitive = caseSensitive;
+    this.batchSize = maxRecordsPerBatch;
+  }
+
+  private static class ReadConf<T> {
+    private final ParquetFileReader reader;
+    private final InputFile file;
+    private final ParquetReadOptions options;
+    private final MessageType projection;
+    private final VectorizedReader<T> model;
+    private final List<BlockMetaData> rowGroups;
+    private final boolean[] shouldSkip;
+    private final long totalValues;
+    private final boolean reuseContainers;
+    private final int batchSize;
+
+    @SuppressWarnings("unchecked")
+    ReadConf(InputFile file, ParquetReadOptions options, Schema expectedSchema, Expression filter,
+        Function<MessageType, VectorizedReader<T>> readerFunc, boolean reuseContainers,
+        boolean caseSensitive, int bSize) {
+      this.file = file;
+      this.options = options;
+      this.reader = newReader(file, options);
+
+      MessageType fileSchema = reader.getFileMetaData().getSchema();
+
+      boolean hasIds = ParquetSchemaUtil.hasIds(fileSchema);
+      MessageType typeWithIds = hasIds ? fileSchema : ParquetSchemaUtil.addFallbackIds(fileSchema);
+
+      this.projection = hasIds ?
+          ParquetSchemaUtil.pruneColumns(fileSchema, expectedSchema) :
+          ParquetSchemaUtil.pruneColumnsFallback(fileSchema, expectedSchema);
+      this.model = readerFunc.apply(typeWithIds);
+      this.rowGroups = reader.getRowGroups();
+      this.shouldSkip = new boolean[rowGroups.size()];
+
+      ParquetMetricsRowGroupFilter statsFilter = null;
+      ParquetDictionaryRowGroupFilter dictFilter = null;
+      if (filter != null) {
+        statsFilter = new ParquetMetricsRowGroupFilter(expectedSchema, filter, caseSensitive);
+        dictFilter = new ParquetDictionaryRowGroupFilter(expectedSchema, filter, caseSensitive);
+      }
+
+      long computedTotalValues = 0L;
+      for (int i = 0; i < shouldSkip.length; i += 1) {
+        BlockMetaData rowGroup = rowGroups.get(i);
+        boolean shouldRead = filter == null || (
+            statsFilter.shouldRead(typeWithIds, rowGroup) &&
+                dictFilter.shouldRead(typeWithIds, rowGroup, reader.getDictionaryReader(rowGroup)));
+        this.shouldSkip[i] = !shouldRead;
+        if (shouldRead) {
+          computedTotalValues += rowGroup.getRowCount();
+        }
+      }
+
+      this.totalValues = computedTotalValues;
+      this.reuseContainers = reuseContainers;
+      this.batchSize = bSize;
+    }
+
+    ReadConf(ReadConf<T> toCopy) {
+      this.reader = null;
+      this.file = toCopy.file;
+      this.options = toCopy.options;
+      this.projection = toCopy.projection;
+      this.model = toCopy.model;
+      this.rowGroups = toCopy.rowGroups;
+      this.shouldSkip = toCopy.shouldSkip;
+      this.totalValues = toCopy.totalValues;
+      this.reuseContainers = toCopy.reuseContainers;
+      this.batchSize = toCopy.batchSize;
+    }
+
+    ParquetFileReader reader() {
+      if (reader != null) {
+        reader.setRequestedSchema(projection);
+        return reader;
+      }
+
+      ParquetFileReader newReader = newReader(file, options);
+      newReader.setRequestedSchema(projection);
+      return newReader;
+    }
+
+    VectorizedReader model() {
+      return model;
+    }
+
+    boolean[] shouldSkip() {
+      return shouldSkip;
+    }
+
+    long totalValues() {
+      return totalValues;
+    }
+
+    boolean reuseContainers() {
+      return reuseContainers;
+    }
+
+    int batchSize() {
+      return batchSize;
+    }
+
+    private static ParquetFileReader newReader(InputFile file, ParquetReadOptions options) {
+      try {
+        return ParquetFileReader.open(ParquetIO.file(file), options);
+      } catch (IOException e) {
+        throw new RuntimeIOException(e, "Failed to open Parquet file: %s", file.location());
+      }
+    }
+
+    ReadConf<T> copy() {
+      return new ReadConf<>(this);
+    }
+  }
+
+  private ReadConf conf = null;
+
+  private ReadConf init() {
+    if (conf == null) {
+      ReadConf readConf = new ReadConf(
+          input, options, expectedSchema, filter, batchReaderFunc, reuseContainers, caseSensitive, batchSize);
+      this.conf = readConf.copy();
+      return readConf;
+    }
+
+    return conf;
+  }
+
+  @Override
+  public Iterator iterator() {
+    FileIterator iter = new FileIterator(init());
+    addCloseable(iter);
+    return iter;
+  }
+
+  private static class FileIterator<T> implements Iterator<T>, Closeable {
+    private final ParquetFileReader reader;
+    private final boolean[] shouldSkip;
+    private final VectorizedReader<T> model;
+    private final long totalValues;
+    private final int batchSize;
+
+    private int nextRowGroup = 0;
+    private long nextRowGroupStart = 0;
+    private long valuesRead = 0;
+    private T last = null;
+
+    FileIterator(ReadConf conf) {
+      this.reader = conf.reader();
+      this.shouldSkip = conf.shouldSkip();
+      this.model = conf.model();
+      this.totalValues = conf.totalValues();
+      this.model.reuseContainers(conf.reuseContainers());
+      this.batchSize = conf.batchSize();
+    }
+
+    @Override
+    public boolean hasNext() {
+      return valuesRead < totalValues;
+    }
+
+    @Override
+    public T next() {
+      if (!hasNext()) {
+        throw new NoSuchElementException();
+      }
+      if (valuesRead >= nextRowGroupStart) {
+        advance();
+      }
+      this.last = model.read();
+      valuesRead += Math.min(nextRowGroupStart - valuesRead, batchSize);
+      return last;
+    }
+
+    private void advance() {
+      while (shouldSkip[nextRowGroup]) {
+        nextRowGroup += 1;
+        reader.skipNextRowGroup();
+      }
+
+      PageReadStore pages;
+      DictionaryPageReadStore dictionaryPageReadStore;
+      try {
+        dictionaryPageReadStore = reader.getNextDictionaryReader();
+        pages = reader.readNextRowGroup();
+      } catch (IOException e) {
+        throw new RuntimeIOException(e);
+      }
+
+      nextRowGroupStart += pages.getRowCount();
+      nextRowGroup += 1;
+      model.setRowGroupInfo(
+          pages,
+          dictionaryPageReadStore,
+          dictionaryPageReadStore == null ? null : buildColumnDictEncodedMap(reader.getRowGroups()));
+    }
+
+    /**
+     * Retuns a map of {@link ColumnPath} -> whether all the pages in the row group for this column are dictionary
+     * encoded
+     */
+    private static Map<ColumnPath, Boolean> buildColumnDictEncodedMap(List<BlockMetaData> blockMetaData) {
+      Map<ColumnPath, Boolean> map = new HashMap<>();
+      for (BlockMetaData b : blockMetaData) {
+        for (ColumnChunkMetaData c : b.getColumns()) {
+          map.put(c.getPath(), !ParquetUtil.hasNonDictionaryPages(c));
+        }
+      }
+      return map;
 
 Review comment:
   Fixed

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


With regards,
Apache Git Services

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