[GitHub] [incubator-pinot] kishoreg commented on a change in pull request #4321: #4317 Feature/variable length bytes offline dictionary for indexing bytes and string dicts.

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

kishoreg commented on a change in pull request #4321: #4317 Feature/variable length bytes offline dictionary for indexing bytes and string dicts.
URL: https://github.com/apache/incubator-pinot/pull/4321#discussion_r295527950
 
 

 ##########
 File path: pinot-core/src/main/java/org/apache/pinot/core/io/util/VarLengthBytesValueReaderWriter.java
 ##########
 @@ -0,0 +1,196 @@
+/**
+ * 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.pinot.core.io.util;
+
+import com.google.common.base.Preconditions;
+import java.io.Closeable;
+import java.io.IOException;
+import java.util.Arrays;
+import org.apache.pinot.common.utils.StringUtil;
+import org.apache.pinot.core.segment.memory.PinotDataBuffer;
+
+/**
+ * Implementation of {@link ValueReader} that will allow each value to be of variable
+ * length and there by avoiding the unnecessary padding.
+ *
+ * The layout of the file is as follows:
+ * <p> Header Section: </p>
+ * <ul>
+ *   <li> Magic header byte sequence. </li>
+ * </ul>
+ *
+ * <p> Values: </p>
+ * <ul>
+ *   <li> Integer offsets to start position of byte arrays. </li>
+ *   <li> All byte arrays. </li>
+ * </ul>
+ *
+ * Only sequential writes are supported.
+ *
+ * @see FixedByteValueReaderWriter
+ */
+public class VarLengthBytesValueReaderWriter implements Closeable, ValueReader {
+
+  /**
+   * Header used to identify the dictionary files written in variable length bytes
+   * format. Keeping this a mix of alphanumeric with special character to avoid
+   * collisions with the regular int/string dictionary values written in fixed size
+   * format.
+   */
+  private static final byte[] MAGIC_HEADER = StringUtil.encodeUtf8("....vl1;");
+  private static final int MAGIC_HEADER_LENGTH = MAGIC_HEADER.length;
+  private static final int INDEX_ARRAY_START_OFFSET = MAGIC_HEADER.length + Integer.BYTES;
+
+  private final PinotDataBuffer _dataBuffer;
+  private transient int _numElements;
+
+  public VarLengthBytesValueReaderWriter(PinotDataBuffer dataBuffer) {
+    this._dataBuffer = dataBuffer;
+  }
+
+  public static long getRequiredSize(byte[][] byteArrays) {
+    // First include the magic header, length field and offsets array.
+    long length = INDEX_ARRAY_START_OFFSET + Integer.BYTES * (byteArrays.length);
+
+    for (byte[] array: byteArrays) {
+      length += array.length;
+    }
+    return length;
+  }
+
+  public static boolean isVarLengthBytesDictBuffer(PinotDataBuffer buffer) {
+    // If the buffer is smaller than header + numElements size, it's not var length dictionary.
+    if (buffer.size() > INDEX_ARRAY_START_OFFSET) {
+      byte[] header = new byte[MAGIC_HEADER_LENGTH];
+      buffer.copyTo(0, header, 0, MAGIC_HEADER_LENGTH);
+
+      if (Arrays.equals(MAGIC_HEADER, header)) {
+        // Also verify that there is a valid numElements value.
+        int numElements = buffer.getInt(MAGIC_HEADER_LENGTH);
+        return numElements > 0;
+      }
+    }
+
+    return false;
+  }
+
+  private void writeHeader() {
+    for (int offset = 0; offset < MAGIC_HEADER_LENGTH; offset++) {
+      _dataBuffer.putByte(offset, MAGIC_HEADER[offset]);
+    }
+  }
+
+  public void init(byte[][] byteArrays) {
+    this._numElements = byteArrays.length;
+
+    writeHeader();
+
+    // Add the number of elements as the first field in the buffer.
+    _dataBuffer.putInt(MAGIC_HEADER_LENGTH, byteArrays.length);
+
+    // Then write the offset of each of the byte array in the data buffer.
+    int nextOffset = INDEX_ARRAY_START_OFFSET;
+    int nextArrayStartOffset = INDEX_ARRAY_START_OFFSET + Integer.BYTES * byteArrays.length;
+    for (byte[] array: byteArrays) {
+      _dataBuffer.putInt(nextOffset, nextArrayStartOffset);
+      nextOffset += Integer.BYTES;
+      nextArrayStartOffset += array.length;
+    }
+
+    // Finally write the byte arrays.
+    nextArrayStartOffset = INDEX_ARRAY_START_OFFSET + Integer.BYTES * byteArrays.length;
+    for (byte[] array: byteArrays) {
+      _dataBuffer.readFrom(nextArrayStartOffset, array);
+      nextArrayStartOffset += array.length;
+    }
+  }
+
+  int getNumElements() {
+    // Lazily initialize the numElements.
+    if (_numElements == 0) {
+      _numElements = _dataBuffer.getInt(MAGIC_HEADER_LENGTH);
+      Preconditions.checkArgument(_numElements > 0);
+    }
+    return this._numElements;
+  }
+
+  @Override
+  public int getInt(int index) {
+    throw new UnsupportedOperationException();
+  }
+
+  @Override
+  public long getLong(int index) {
+    throw new UnsupportedOperationException();
+  }
+
+  @Override
+  public float getFloat(int index) {
+    throw new UnsupportedOperationException();
+  }
+
+  @Override
+  public double getDouble(int index) {
+    throw new UnsupportedOperationException();
+  }
+
+  @Override
+  public String getUnpaddedString(int index, int numBytesPerValue, byte paddingByte, byte[] buffer) {
+    return StringUtil.decodeUtf8(getBytes(index, numBytesPerValue, buffer));
+  }
+
+  @Override
+  public String getPaddedString(int index, int numBytesPerValue, byte[] buffer) {
+    return StringUtil.decodeUtf8(getBytes(index, numBytesPerValue, buffer));
+  }
+
+  @Override
+  public byte[] getBytes(int index, int numBytesPerValue, byte[] buffer) {
 
 Review comment:
   can we store an additional offset to make this code very simple without any branching?
   length = buf[i+1] - buf[i]

----------------------------------------------------------------
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: commits-unsubscribe@pinot.apache.org
For additional commands, e-mail: commits-help@pinot.apache.org