[GitHub] carbondata pull request #2437: [WIP] segregate block and blocklet cache

[qiuchenjian-2]

GitHub user manishgupta88 opened a pull request:

    https://github.com/apache/carbondata/pull/2437

    [WIP] segregate block and blocklet cache

    segregate block and blocklet cache
   
     - [ ] Any interfaces changed?
     
     - [ ] Any backward compatibility impacted?
     
     - [ ] Document update required?
   
     - [ ] Testing done
            Please provide details on
            - Whether new unit test cases have been added or why no new tests are required?
            - How it is tested? Please attach test report.
            - Is it a performance related change? Please attach the performance test report.
            - Any additional information to help reviewers in testing this change.
           
     - [ ] For large changes, please consider breaking it into sub-tasks under an umbrella JIRA.
   


You can merge this pull request into a Git repository by running:

    $ git pull https://github.com/manishgupta88/carbondata block_blocklet_cache

Alternatively you can review and apply these changes as the patch at:

    https://github.com/apache/carbondata/pull/2437.patch

To close this pull request, make a commit to your master/trunk branch
with (at least) the following in the commit message:

    This closes #2437
   
----
commit 8c8215da55a687f4debadd56fa9d5356729d3331
Author: manishgupta88 <tomanishgupta18@...>
Date:   2018-06-25T06:43:00Z

    segregate block and blocklet cache

----


---

[qiuchenjian-2]

Github user CarbonDataQA commented on the issue:

    https://github.com/apache/carbondata/pull/2437
 
    Build Failed  with Spark 2.1.0, Please check CI http://136.243.101.176:8080/job/ApacheCarbonPRBuilder1/6718/



---

[qiuchenjian-2]

Github user CarbonDataQA commented on the issue:

    https://github.com/apache/carbondata/pull/2437
 
    Build Failed with Spark 2.2.1, Please check CI http://88.99.58.216:8080/job/ApacheCarbonPRBuilder/5545/



---

[qiuchenjian-2]

Github user ravipesala commented on the issue:

    https://github.com/apache/carbondata/pull/2437
 
    SDV Build Fail , Please check CI http://144.76.159.231:8080/job/ApacheSDVTests/5567/



---

[qiuchenjian-2]

Github user CarbonDataQA commented on the issue:

    https://github.com/apache/carbondata/pull/2437
 
    Build Failed with Spark 2.2.1, Please check CI http://88.99.58.216:8080/job/ApacheCarbonPRBuilder/5558/



---

[qiuchenjian-2]

Github user CarbonDataQA commented on the issue:

    https://github.com/apache/carbondata/pull/2437
 
    Build Failed  with Spark 2.1.0, Please check CI http://136.243.101.176:8080/job/ApacheCarbonPRBuilder1/6728/



---

[qiuchenjian-2]

Github user CarbonDataQA commented on the issue:

    https://github.com/apache/carbondata/pull/2437
 
    Build Failed with Spark 2.2.1, Please check CI http://88.99.58.216:8080/job/ApacheCarbonPRBuilder/5568/



---

[qiuchenjian-2]

Github user ravipesala commented on the issue:

    https://github.com/apache/carbondata/pull/2437
 
    SDV Build Fail , Please check CI http://144.76.159.231:8080/job/ApacheSDVTests/5580/



---

[qiuchenjian-2]

Github user CarbonDataQA commented on the issue:

    https://github.com/apache/carbondata/pull/2437
 
    Build Failed  with Spark 2.1.0, Please check CI http://136.243.101.176:8080/job/ApacheCarbonPRBuilder1/6738/



---

[qiuchenjian-2]

Github user ravipesala commented on a diff in the pull request:

    https://github.com/apache/carbondata/pull/2437#discussion_r200149538
 
    --- Diff: core/src/main/java/org/apache/carbondata/core/indexstore/row/DataMapRow.java ---
    @@ -26,70 +26,66 @@
      */
     public abstract class DataMapRow implements Serializable {
     
    -  protected CarbonRowSchema[] schemas;
    --- End diff --
   
    I don't think it is required to remove the schema from here as it is a temporary object to read data. It does not occupy any heap memory.  Many changes related to this is not required. Please revert the changes


---

[qiuchenjian-2]

Github user ravipesala commented on a diff in the pull request:

    https://github.com/apache/carbondata/pull/2437#discussion_r200149739
 
    --- Diff: core/src/main/java/org/apache/carbondata/core/indexstore/blockletindex/BlockDataMap.java ---
    @@ -0,0 +1,879 @@
    +/*
    + * 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.carbondata.core.indexstore.blockletindex;
    +
    +import java.io.ByteArrayInputStream;
    +import java.io.ByteArrayOutputStream;
    +import java.io.DataInput;
    +import java.io.DataInputStream;
    +import java.io.DataOutput;
    +import java.io.DataOutputStream;
    +import java.io.IOException;
    +import java.io.Serializable;
    +import java.io.UnsupportedEncodingException;
    --- End diff --
   
    use `import java.io.*`


---

[qiuchenjian-2]

Github user ravipesala commented on a diff in the pull request:

    https://github.com/apache/carbondata/pull/2437#discussion_r200155521
 
    --- Diff: core/src/main/java/org/apache/carbondata/core/indexstore/blockletindex/BlockDataMap.java ---
    @@ -0,0 +1,879 @@
    +/*
    + * 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.carbondata.core.indexstore.blockletindex;
    +
    +import java.io.ByteArrayInputStream;
    +import java.io.ByteArrayOutputStream;
    +import java.io.DataInput;
    +import java.io.DataInputStream;
    +import java.io.DataOutput;
    +import java.io.DataOutputStream;
    +import java.io.IOException;
    +import java.io.Serializable;
    +import java.io.UnsupportedEncodingException;
    +import java.util.ArrayList;
    +import java.util.BitSet;
    +import java.util.List;
    +
    +import org.apache.carbondata.common.logging.LogService;
    +import org.apache.carbondata.common.logging.LogServiceFactory;
    +import org.apache.carbondata.core.constants.CarbonCommonConstants;
    +import org.apache.carbondata.core.datamap.dev.DataMapModel;
    +import org.apache.carbondata.core.datamap.dev.cgdatamap.CoarseGrainDataMap;
    +import org.apache.carbondata.core.datastore.block.SegmentProperties;
    +import org.apache.carbondata.core.datastore.block.TableBlockInfo;
    +import org.apache.carbondata.core.indexstore.AbstractMemoryDMStore;
    +import org.apache.carbondata.core.indexstore.BlockMetaInfo;
    +import org.apache.carbondata.core.indexstore.Blocklet;
    +import org.apache.carbondata.core.indexstore.BlockletDetailInfo;
    +import org.apache.carbondata.core.indexstore.ExtendedBlocklet;
    +import org.apache.carbondata.core.indexstore.PartitionSpec;
    +import org.apache.carbondata.core.indexstore.SafeMemoryDMStore;
    +import org.apache.carbondata.core.indexstore.UnsafeMemoryDMStore;
    +import org.apache.carbondata.core.indexstore.row.DataMapRow;
    +import org.apache.carbondata.core.indexstore.row.DataMapRowImpl;
    +import org.apache.carbondata.core.indexstore.schema.CarbonRowSchema;
    +import org.apache.carbondata.core.indexstore.schema.SchemaGenerator;
    +import org.apache.carbondata.core.memory.MemoryException;
    +import org.apache.carbondata.core.metadata.blocklet.DataFileFooter;
    +import org.apache.carbondata.core.metadata.blocklet.index.BlockletIndex;
    +import org.apache.carbondata.core.metadata.blocklet.index.BlockletMinMaxIndex;
    +import org.apache.carbondata.core.metadata.schema.table.column.ColumnSchema;
    +import org.apache.carbondata.core.profiler.ExplainCollector;
    +import org.apache.carbondata.core.scan.filter.FilterExpressionProcessor;
    +import org.apache.carbondata.core.scan.filter.FilterUtil;
    +import org.apache.carbondata.core.scan.filter.executer.FilterExecuter;
    +import org.apache.carbondata.core.scan.filter.executer.ImplicitColumnFilterExecutor;
    +import org.apache.carbondata.core.scan.filter.resolver.FilterResolverIntf;
    +import org.apache.carbondata.core.util.BlockletDataMapUtil;
    +import org.apache.carbondata.core.util.ByteUtil;
    +import org.apache.carbondata.core.util.CarbonUtil;
    +import org.apache.carbondata.core.util.DataFileFooterConverter;
    +import org.apache.carbondata.core.util.path.CarbonTablePath;
    +
    +import org.apache.commons.lang3.ArrayUtils;
    +import org.apache.commons.lang3.StringUtils;
    +import org.apache.hadoop.fs.Path;
    +import org.xerial.snappy.Snappy;
    +
    +/**
    + * Datamap implementation for blocklet.
    + */
    +public class BlockDataMap extends CoarseGrainDataMap implements Serializable {
    +
    +  private static final LogService LOGGER =
    +      LogServiceFactory.getLogService(BlockDataMap.class.getName());
    +
    +  protected static final long serialVersionUID = -2170289352240810993L;
    +  /**
    +   * for CACHE_LEVEL=BLOCK and legacy store default blocklet id will be -1
    +   */
    +  private static final short BLOCK_DEFAULT_BLOCKLET_ID = -1;
    +
    +  protected AbstractMemoryDMStore memoryDMStore;
    +
    +  protected AbstractMemoryDMStore taskSummaryDMStore;
    +
    +  // As it is a heavy object it is not recommended to serialize this object
    +  protected transient SegmentProperties segmentProperties;
    +
    +  protected int[] columnCardinality;
    +
    +  protected long blockletSchemaTime;
    +  /**
    +   * flag to check for store from 1.1 or any prior version
    +   */
    +  protected boolean isLegacyStore;
    +
    +  @Override public void init(DataMapModel dataMapModel) throws IOException, MemoryException {
    +    long startTime = System.currentTimeMillis();
    +    assert (dataMapModel instanceof BlockletDataMapModel);
    +    BlockletDataMapModel blockletDataMapInfo = (BlockletDataMapModel) dataMapModel;
    +    DataFileFooterConverter fileFooterConverter = new DataFileFooterConverter();
    +    List<DataFileFooter> indexInfo = fileFooterConverter
    +        .getIndexInfo(blockletDataMapInfo.getFilePath(), blockletDataMapInfo.getFileData());
    +    Path path = new Path(blockletDataMapInfo.getFilePath());
    +    byte[] filePath = path.getParent().toString().getBytes(CarbonCommonConstants.DEFAULT_CHARSET);
    +    byte[] fileName = path.getName().toString().getBytes(CarbonCommonConstants.DEFAULT_CHARSET);
    +    byte[] segmentId =
    +        blockletDataMapInfo.getSegmentId().getBytes(CarbonCommonConstants.DEFAULT_CHARSET);
    +    byte[] schemaBinary = null;
    +    if (!indexInfo.isEmpty()) {
    +      DataFileFooter fileFooter = indexInfo.get(0);
    +      // store for 1.1 or any prior version will not have any blocklet information in file footer
    +      isLegacyStore = fileFooter.getBlockletList() == null;
    +      // init segment properties and create schema
    +      initSegmentProperties(fileFooter);
    +      schemaBinary = convertSchemaToBinary(fileFooter.getColumnInTable());
    +      createSchema(segmentProperties, blockletDataMapInfo.isAddToUnsafe(), true);
    +      createSummarySchema(segmentProperties, schemaBinary, filePath, fileName, segmentId,
    +          blockletDataMapInfo.isAddToUnsafe(), true);
    +    }
    +    // check for legacy store and load the metadata
    +    DataMapRowImpl summaryRow = loadBlockMetadata(blockletDataMapInfo, indexInfo);
    --- End diff --
   
    Create protected methods for `createSchema`, `createSummarySchema` and `loadMetadata` and override the same in BloclkletDataMap


---

[qiuchenjian-2]

Github user ravipesala commented on a diff in the pull request:

    https://github.com/apache/carbondata/pull/2437#discussion_r200156093
 
    --- Diff: core/src/main/java/org/apache/carbondata/core/indexstore/blockletindex/BlockletDataMap.java ---
    @@ -221,438 +159,71 @@ private DataMapRowImpl loadToUnsafe(DataFileFooter fileFooter,
         CarbonRowSchema[] schema = memoryDMStore.getSchema();
         // Add one row to maintain task level min max for segment pruning
         if (!blockletList.isEmpty() && summaryRow == null) {
    -      summaryRow = new DataMapRowImpl(summaryDMStore.getSchema());
    +      summaryRow = new DataMapRowImpl(taskSummaryDMStore.getSchema().length);
         }
         for (int index = 0; index < blockletList.size(); index++) {
    -      DataMapRow row = new DataMapRowImpl(schema);
    +      DataMapRow row = new DataMapRowImpl(schema.length);
           int ordinal = 0;
           int taskMinMaxOrdinal = 0;
           BlockletInfo blockletInfo = blockletList.get(index);
    -
    -      // add start key as index key
    -      row.setByteArray(blockletInfo.getBlockletIndex().getBtreeIndex().getStartKey(), ordinal++);
    -
           BlockletMinMaxIndex minMaxIndex = blockletInfo.getBlockletIndex().getMinMaxIndex();
    -      byte[][] minValues = updateMinValues(minMaxIndex.getMinValues(), minMaxLen);
    -      row.setRow(addMinMax(minMaxLen, schema[ordinal], minValues), ordinal);
    +      row.setRow(addMinMax(minMaxLen, schema[ordinal], minMaxIndex.getMinValues()), ordinal);
           // compute and set task level min values
    -      addTaskMinMaxValues(summaryRow, minMaxLen,
    -          summaryDMStore.getSchema()[taskMinMaxOrdinal], minValues,
    -          TASK_MIN_VALUES_INDEX, true);
    +      addTaskMinMaxValues(summaryRow, minMaxLen, taskSummaryDMStore.getSchema(), taskMinMaxOrdinal,
    +          minMaxIndex.getMinValues(), BlockletDataMapRowIndexes.TASK_MIN_VALUES_INDEX, true);
           ordinal++;
           taskMinMaxOrdinal++;
    -      byte[][] maxValues = updateMaxValues(minMaxIndex.getMaxValues(), minMaxLen);
    -      row.setRow(addMinMax(minMaxLen, schema[ordinal], maxValues), ordinal);
    +      row.setRow(addMinMax(minMaxLen, schema[ordinal], minMaxIndex.getMaxValues()), ordinal);
           // compute and set task level max values
    -      addTaskMinMaxValues(summaryRow, minMaxLen,
    -          summaryDMStore.getSchema()[taskMinMaxOrdinal], maxValues,
    -          TASK_MAX_VALUES_INDEX, false);
    +      addTaskMinMaxValues(summaryRow, minMaxLen, taskSummaryDMStore.getSchema(), taskMinMaxOrdinal,
    +          minMaxIndex.getMaxValues(), BlockletDataMapRowIndexes.TASK_MAX_VALUES_INDEX, false);
           ordinal++;
    -
           row.setInt(blockletInfo.getNumberOfRows(), ordinal++);
    -
           // add file path
           byte[] filePathBytes = filePath.getBytes(CarbonCommonConstants.DEFAULT_CHARSET_CLASS);
           row.setByteArray(filePathBytes, ordinal++);
    -
    -      // add pages
    -      row.setShort((short) blockletInfo.getNumberOfPages(), ordinal++);
    -
           // add version number
           row.setShort(fileFooter.getVersionId().number(), ordinal++);
    -
           // add schema updated time
           row.setLong(fileFooter.getSchemaUpdatedTimeStamp(), ordinal++);
    -
    -      // add blocklet info
           byte[] serializedData;
           try {
    +        // Add block footer offset, it is used if we need to read footer of block
    +        row.setLong(fileFooter.getBlockInfo().getTableBlockInfo().getBlockOffset(), ordinal++);
    +        setLocations(blockMetaInfo.getLocationInfo(), row, ordinal++);
    +        // Store block size
    +        row.setLong(blockMetaInfo.getSize(), ordinal++);
    +        // add blocklet info
             ByteArrayOutputStream stream = new ByteArrayOutputStream();
             DataOutput dataOutput = new DataOutputStream(stream);
             blockletInfo.write(dataOutput);
             serializedData = stream.toByteArray();
             row.setByteArray(serializedData, ordinal++);
    -        // Add block footer offset, it is used if we need to read footer of block
    -        row.setLong(fileFooter.getBlockInfo().getTableBlockInfo().getBlockOffset(), ordinal++);
    -        setLocations(blockMetaInfo.getLocationInfo(), row, ordinal);
    -        ordinal++;
    -        // for relative blockelt id i.e blocklet id that belongs to a particular part file
    -        row.setShort((short) relativeBlockletId++, ordinal++);
    -        // Store block size
    -        row.setLong(blockMetaInfo.getSize(), ordinal);
    +        // add pages
    +        row.setShort((short) blockletInfo.getNumberOfPages(), ordinal++);
    +        // for relative blocklet id i.e blocklet id that belongs to a particular carbondata file
    +        row.setShort((short) relativeBlockletId++, ordinal);
             memoryDMStore.addIndexRow(row);
           } catch (Exception e) {
             throw new RuntimeException(e);
           }
         }
    -
         return summaryRow;
       }
     
    -  private void setLocations(String[] locations, DataMapRow row, int ordinal)
    -      throws UnsupportedEncodingException {
    -    // Add location info
    -    String locationStr = StringUtils.join(locations, ',');
    -    row.setByteArray(locationStr.getBytes(CarbonCommonConstants.DEFAULT_CHARSET), ordinal);
    -  }
    -
    -  /**
    -   * Load information for the block.It is the case can happen only for old stores
    -   * where blocklet information is not available in index file. So load only block information
    -   * and read blocklet information in executor.
    -   */
    -  private DataMapRowImpl loadToUnsafeBlock(DataFileFooter fileFooter,
    -      SegmentProperties segmentProperties, String filePath, DataMapRowImpl summaryRow,
    -      BlockMetaInfo blockMetaInfo) {
    -    int[] minMaxLen = segmentProperties.getColumnsValueSize();
    -    BlockletIndex blockletIndex = fileFooter.getBlockletIndex();
    -    CarbonRowSchema[] schema = memoryDMStore.getSchema();
    -    // Add one row to maintain task level min max for segment pruning
    -    if (summaryRow == null) {
    -      summaryRow = new DataMapRowImpl(summaryDMStore.getSchema());
    -    }
    -    DataMapRow row = new DataMapRowImpl(schema);
    -    int ordinal = 0;
    -    int taskMinMaxOrdinal = 0;
    -    // add start key as index key
    -    row.setByteArray(blockletIndex.getBtreeIndex().getStartKey(), ordinal++);
    -
    -    BlockletMinMaxIndex minMaxIndex = blockletIndex.getMinMaxIndex();
    -    byte[][] minValues = updateMinValues(minMaxIndex.getMinValues(), minMaxLen);
    -    byte[][] maxValues = updateMaxValues(minMaxIndex.getMaxValues(), minMaxLen);
    -    // update min max values in case of old store
    -    byte[][] updatedMinValues =
    -        CarbonUtil.updateMinMaxValues(fileFooter, maxValues, minValues, true);
    -    byte[][] updatedMaxValues =
    -        CarbonUtil.updateMinMaxValues(fileFooter, maxValues, minValues, false);
    -    row.setRow(addMinMax(minMaxLen, schema[ordinal], updatedMinValues), ordinal);
    -    // compute and set task level min values
    -    addTaskMinMaxValues(summaryRow, minMaxLen,
    -        summaryDMStore.getSchema()[taskMinMaxOrdinal], updatedMinValues,
    -        TASK_MIN_VALUES_INDEX, true);
    -    ordinal++;
    -    taskMinMaxOrdinal++;
    -    row.setRow(addMinMax(minMaxLen, schema[ordinal], updatedMaxValues), ordinal);
    -    // compute and set task level max values
    -    addTaskMinMaxValues(summaryRow, minMaxLen,
    -        summaryDMStore.getSchema()[taskMinMaxOrdinal], updatedMaxValues,
    -        TASK_MAX_VALUES_INDEX, false);
    -    ordinal++;
    -
    -    row.setInt((int)fileFooter.getNumberOfRows(), ordinal++);
    -
    -    // add file path
    -    byte[] filePathBytes = filePath.getBytes(CarbonCommonConstants.DEFAULT_CHARSET_CLASS);
    -    row.setByteArray(filePathBytes, ordinal++);
    -
    -    // add pages
    -    row.setShort((short) 0, ordinal++);
    -
    -    // add version number
    -    row.setShort(fileFooter.getVersionId().number(), ordinal++);
    -
    -    // add schema updated time
    -    row.setLong(fileFooter.getSchemaUpdatedTimeStamp(), ordinal++);
    -
    -    // add blocklet info
    -    row.setByteArray(new byte[0], ordinal++);
    -
    -    row.setLong(fileFooter.getBlockInfo().getTableBlockInfo().getBlockOffset(), ordinal++);
    -    try {
    -      setLocations(blockMetaInfo.getLocationInfo(), row, ordinal);
    -      ordinal++;
    -      // for relative blocklet id. Value is -1 because in case of old store blocklet info will
    -      // not be present in the index file and in that case we will not knwo the total number of
    -      // blocklets
    -      row.setShort((short) -1, ordinal++);
    -
    -      // store block size
    -      row.setLong(blockMetaInfo.getSize(), ordinal);
    -      memoryDMStore.addIndexRow(row);
    -    } catch (Exception e) {
    -      throw new RuntimeException(e);
    -    }
    -
    -    return summaryRow;
    -  }
    -
    -  private void addTaskSummaryRowToUnsafeMemoryStore(DataMapRow summaryRow, byte[] schemaBinary,
    -      byte[] filePath, byte[] fileName, byte[] segmentId) {
    -    // write the task summary info to unsafe memory store
    -    if (null != summaryRow) {
    -      // Add column schema , it is useful to generate segment properties in executor.
    -      // So we no need to read footer again there.
    -      if (schemaBinary != null) {
    -        summaryRow.setByteArray(schemaBinary, SCHEMA);
    -      }
    -      summaryRow.setByteArray(filePath, INDEX_PATH);
    -      summaryRow.setByteArray(fileName, INDEX_FILE_NAME);
    -      summaryRow.setByteArray(segmentId, SEGMENTID);
    -      try {
    -        summaryDMStore.addIndexRow(summaryRow);
    -      } catch (Exception e) {
    -        throw new RuntimeException(e);
    -      }
    -    }
    -  }
    -
    -  /**
    -   * Fill the measures min values with minimum , this is needed for backward version compatability
    -   * as older versions don't store min values for measures
    -   */
    -  private byte[][] updateMinValues(byte[][] minValues, int[] minMaxLen) {
    -    byte[][] updatedValues = minValues;
    -    if (minValues.length < minMaxLen.length) {
    -      updatedValues = new byte[minMaxLen.length][];
    -      System.arraycopy(minValues, 0, updatedValues, 0, minValues.length);
    -      List<CarbonMeasure> measures = segmentProperties.getMeasures();
    -      ByteBuffer buffer = ByteBuffer.allocate(8);
    -      for (int i = 0; i < measures.size(); i++) {
    -        buffer.rewind();
    -        DataType dataType = measures.get(i).getDataType();
    -        if (dataType == DataTypes.BYTE) {
    -          buffer.putLong(Byte.MIN_VALUE);
    -          updatedValues[minValues.length + i] = buffer.array().clone();
    -        } else if (dataType == DataTypes.SHORT) {
    -          buffer.putLong(Short.MIN_VALUE);
    -          updatedValues[minValues.length + i] = buffer.array().clone();
    -        } else if (dataType == DataTypes.INT) {
    -          buffer.putLong(Integer.MIN_VALUE);
    -          updatedValues[minValues.length + i] = buffer.array().clone();
    -        } else if (dataType == DataTypes.LONG) {
    -          buffer.putLong(Long.MIN_VALUE);
    -          updatedValues[minValues.length + i] = buffer.array().clone();
    -        } else if (DataTypes.isDecimal(dataType)) {
    -          updatedValues[minValues.length + i] =
    -              DataTypeUtil.bigDecimalToByte(BigDecimal.valueOf(Long.MIN_VALUE));
    -        } else {
    -          buffer.putDouble(Double.MIN_VALUE);
    -          updatedValues[minValues.length + i] = buffer.array().clone();
    -        }
    -      }
    -    }
    -    return updatedValues;
    -  }
    -
    -  /**
    -   * Fill the measures max values with maximum , this is needed for backward version compatability
    -   * as older versions don't store max values for measures
    -   */
    -  private byte[][] updateMaxValues(byte[][] maxValues, int[] minMaxLen) {
    -    byte[][] updatedValues = maxValues;
    -    if (maxValues.length < minMaxLen.length) {
    -      updatedValues = new byte[minMaxLen.length][];
    -      System.arraycopy(maxValues, 0, updatedValues, 0, maxValues.length);
    -      List<CarbonMeasure> measures = segmentProperties.getMeasures();
    -      ByteBuffer buffer = ByteBuffer.allocate(8);
    -      for (int i = 0; i < measures.size(); i++) {
    -        buffer.rewind();
    -        DataType dataType = measures.get(i).getDataType();
    -        if (dataType == DataTypes.BYTE) {
    -          buffer.putLong(Byte.MAX_VALUE);
    -          updatedValues[maxValues.length + i] = buffer.array().clone();
    -        } else if (dataType == DataTypes.SHORT) {
    -          buffer.putLong(Short.MAX_VALUE);
    -          updatedValues[maxValues.length + i] = buffer.array().clone();
    -        } else if (dataType == DataTypes.INT) {
    -          buffer.putLong(Integer.MAX_VALUE);
    -          updatedValues[maxValues.length + i] = buffer.array().clone();
    -        } else if (dataType == DataTypes.LONG) {
    -          buffer.putLong(Long.MAX_VALUE);
    -          updatedValues[maxValues.length + i] = buffer.array().clone();
    -        } else if (DataTypes.isDecimal(dataType)) {
    -          updatedValues[maxValues.length + i] =
    -              DataTypeUtil.bigDecimalToByte(BigDecimal.valueOf(Long.MAX_VALUE));
    -        } else {
    -          buffer.putDouble(Double.MAX_VALUE);
    -          updatedValues[maxValues.length + i] = buffer.array().clone();
    -        }
    -      }
    -    }
    -    return updatedValues;
    -  }
    -
    -  private DataMapRow addMinMax(int[] minMaxLen, CarbonRowSchema carbonRowSchema,
    -      byte[][] minValues) {
    -    CarbonRowSchema[] minSchemas =
    -        ((CarbonRowSchema.StructCarbonRowSchema) carbonRowSchema).getChildSchemas();
    -    DataMapRow minRow = new DataMapRowImpl(minSchemas);
    -    int minOrdinal = 0;
    -    // min value adding
    -    for (int i = 0; i < minMaxLen.length; i++) {
    -      minRow.setByteArray(minValues[i], minOrdinal++);
    -    }
    -    return minRow;
    -  }
    -
    -  /**
    -   * This method will compute min/max values at task level
    -   *
    -   * @param taskMinMaxRow
    -   * @param minMaxLen
    -   * @param carbonRowSchema
    -   * @param minMaxValue
    -   * @param ordinal
    -   * @param isMinValueComparison
    -   */
    -  private void addTaskMinMaxValues(DataMapRow taskMinMaxRow, int[] minMaxLen,
    -      CarbonRowSchema carbonRowSchema, byte[][] minMaxValue, int ordinal,
    -      boolean isMinValueComparison) {
    -    DataMapRow row = taskMinMaxRow.getRow(ordinal);
    -    byte[][] updatedMinMaxValues = minMaxValue;
    -    if (null == row) {
    -      CarbonRowSchema[] minSchemas =
    -          ((CarbonRowSchema.StructCarbonRowSchema) carbonRowSchema).getChildSchemas();
    -      row = new DataMapRowImpl(minSchemas);
    -    } else {
    -      byte[][] existingMinMaxValues = getMinMaxValue(taskMinMaxRow, ordinal);
    -      // Compare and update min max values
    -      for (int i = 0; i < minMaxLen.length; i++) {
    -        int compare =
    -            ByteUtil.UnsafeComparer.INSTANCE.compareTo(existingMinMaxValues[i], minMaxValue[i]);
    -        if (isMinValueComparison) {
    -          if (compare < 0) {
    -            updatedMinMaxValues[i] = existingMinMaxValues[i];
    -          }
    -        } else if (compare > 0) {
    -          updatedMinMaxValues[i] = existingMinMaxValues[i];
    -        }
    -      }
    -    }
    -    int minMaxOrdinal = 0;
    -    // min/max value adding
    -    for (int i = 0; i < minMaxLen.length; i++) {
    -      row.setByteArray(updatedMinMaxValues[i], minMaxOrdinal++);
    -    }
    -    taskMinMaxRow.setRow(row, ordinal);
    -  }
    -
    -  private void createSchema(SegmentProperties segmentProperties, boolean addToUnsafe)
    -      throws MemoryException {
    -    List<CarbonRowSchema> indexSchemas = new ArrayList<>();
    -
    -    // Index key
    -    indexSchemas.add(new CarbonRowSchema.VariableCarbonRowSchema(DataTypes.BYTE_ARRAY));
    -    getMinMaxSchema(segmentProperties, indexSchemas);
    -
    -    // for number of rows.
    -    indexSchemas.add(new CarbonRowSchema.FixedCarbonRowSchema(DataTypes.INT));
    -
    -    // for table block path
    -    indexSchemas.add(new CarbonRowSchema.VariableCarbonRowSchema(DataTypes.BYTE_ARRAY));
    -
    -    // for number of pages.
    -    indexSchemas.add(new CarbonRowSchema.FixedCarbonRowSchema(DataTypes.SHORT));
    -
    -    // for version number.
    -    indexSchemas.add(new CarbonRowSchema.FixedCarbonRowSchema(DataTypes.SHORT));
    -
    -    // for schema updated time.
    -    indexSchemas.add(new CarbonRowSchema.FixedCarbonRowSchema(DataTypes.LONG));
    -
    -    //for blocklet info
    -    indexSchemas.add(new CarbonRowSchema.VariableCarbonRowSchema(DataTypes.BYTE_ARRAY));
    -
    -    // for block footer offset.
    -    indexSchemas.add(new CarbonRowSchema.FixedCarbonRowSchema(DataTypes.LONG));
    -
    -    // for locations
    -    indexSchemas.add(new CarbonRowSchema.VariableCarbonRowSchema(DataTypes.BYTE_ARRAY));
    -
    -    // for relative blocklet id i.e. blocklet id that belongs to a particular part file.
    -    indexSchemas.add(new CarbonRowSchema.FixedCarbonRowSchema(DataTypes.SHORT));
    -
    -    // for storing block length.
    -    indexSchemas.add(new CarbonRowSchema.FixedCarbonRowSchema(DataTypes.LONG));
    -
    -    CarbonRowSchema[] schema = indexSchemas.toArray(new CarbonRowSchema[indexSchemas.size()]);
    -    memoryDMStore = getMemoryDMStore(schema, addToUnsafe);
    -  }
    -
    -  /**
    -   * Creates the schema to store summary information or the information which can be stored only
    -   * once per datamap. It stores datamap level max/min of each column and partition information of
    -   * datamap
    -   * @param segmentProperties
    -   * @throws MemoryException
    -   */
    -  private void createSummarySchema(SegmentProperties segmentProperties, byte[] schemaBinary,
    -      byte[] filePath, byte[] fileName, byte[] segmentId, boolean addToUnsafe)
    -      throws MemoryException {
    -    List<CarbonRowSchema> taskMinMaxSchemas = new ArrayList<>();
    -    getMinMaxSchema(segmentProperties, taskMinMaxSchemas);
    -    // for storing column schema
    -    taskMinMaxSchemas.add(
    -        new CarbonRowSchema.FixedCarbonRowSchema(DataTypes.BYTE_ARRAY, schemaBinary.length));
    -    // for storing file path
    -    taskMinMaxSchemas.add(
    -        new CarbonRowSchema.FixedCarbonRowSchema(DataTypes.BYTE_ARRAY, filePath.length));
    -    // for storing file name
    -    taskMinMaxSchemas.add(
    -        new CarbonRowSchema.FixedCarbonRowSchema(DataTypes.BYTE_ARRAY, fileName.length));
    -    // for storing segmentid
    -    taskMinMaxSchemas.add(
    -        new CarbonRowSchema.FixedCarbonRowSchema(DataTypes.BYTE_ARRAY, segmentId.length));
    -    CarbonRowSchema[] schema =
    -        taskMinMaxSchemas.toArray(new CarbonRowSchema[taskMinMaxSchemas.size()]);
    -    summaryDMStore = getMemoryDMStore(schema, addToUnsafe);
    -  }
    -
    -  private void getMinMaxSchema(SegmentProperties segmentProperties,
    -      List<CarbonRowSchema> minMaxSchemas) {
    -    // Index key
    -    int[] minMaxLen = segmentProperties.getColumnsValueSize();
    -    // do it 2 times, one for min and one for max.
    -    for (int k = 0; k < 2; k++) {
    -      CarbonRowSchema[] mapSchemas = new CarbonRowSchema[minMaxLen.length];
    -      for (int i = 0; i < minMaxLen.length; i++) {
    -        if (minMaxLen[i] <= 0) {
    -          boolean isVarchar = false;
    -          if (i < segmentProperties.getDimensions().size()
    -              && segmentProperties.getDimensions().get(i).getDataType() == DataTypes.VARCHAR) {
    -            isVarchar = true;
    -          }
    -          mapSchemas[i] = new CarbonRowSchema.VariableCarbonRowSchema(DataTypes.BYTE_ARRAY,
    -              isVarchar);
    -        } else {
    -          mapSchemas[i] =
    -              new CarbonRowSchema.FixedCarbonRowSchema(DataTypes.BYTE_ARRAY, minMaxLen[i]);
    -        }
    -      }
    -      CarbonRowSchema mapSchema =
    -          new CarbonRowSchema.StructCarbonRowSchema(DataTypes.createDefaultStructType(),
    -              mapSchemas);
    -      minMaxSchemas.add(mapSchema);
    -    }
    -  }
    -
    -  @Override
    -  public boolean isScanRequired(FilterResolverIntf filterExp) {
    -    FilterExecuter filterExecuter =
    -        FilterUtil.getFilterExecuterTree(filterExp, segmentProperties, null);
    -    for (int i = 0; i < summaryDMStore.getRowCount(); i++) {
    -      DataMapRow unsafeRow = summaryDMStore.getDataMapRow(i);
    -      boolean isScanRequired = FilterExpressionProcessor.isScanRequired(
    -          filterExecuter, getMinMaxValue(unsafeRow, TASK_MAX_VALUES_INDEX),
    -          getMinMaxValue(unsafeRow, TASK_MIN_VALUES_INDEX));
    -      if (isScanRequired) {
    -        return true;
    -      }
    -    }
    -    return false;
    -  }
    -
       private List<Blocklet> prune(FilterResolverIntf filterExp, SegmentProperties segmentProperties) {
         if (memoryDMStore.getRowCount() == 0) {
           return new ArrayList<>();
         }
         List<Blocklet> blocklets = new ArrayList<>();
    +    CarbonRowSchema[] schema = memoryDMStore.getSchema();
    --- End diff --
   
    Use the prune from super class


---

[qiuchenjian-2]

Github user ravipesala commented on a diff in the pull request:

    https://github.com/apache/carbondata/pull/2437#discussion_r200157458
 
    --- Diff: core/src/main/java/org/apache/carbondata/core/indexstore/blockletindex/BlockletDataMapRowIndexes.java ---
    @@ -0,0 +1,65 @@
    +/*
    + * 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.carbondata.core.indexstore.blockletindex;
    +
    +/**
    + * holder for blocklet info indexes in a DataMap row
    + */
    +public class BlockletDataMapRowIndexes {
    --- End diff --
   
    Change to interface


---

[qiuchenjian-2]

Github user ravipesala commented on the issue:

    https://github.com/apache/carbondata/pull/2437
 
    SDV Build Success , Please check CI http://144.76.159.231:8080/job/ApacheSDVTests/5608/



---

[qiuchenjian-2]

Github user ravipesala commented on the issue:

    https://github.com/apache/carbondata/pull/2437
 
    SDV Build Success , Please check CI http://144.76.159.231:8080/job/ApacheSDVTests/5611/



---

[qiuchenjian-2]

Github user CarbonDataQA commented on the issue:

    https://github.com/apache/carbondata/pull/2437
 
    Build Success with Spark 2.2.1, Please check CI http://88.99.58.216:8080/job/ApacheCarbonPRBuilder/5598/



---

[qiuchenjian-2]

Github user CarbonDataQA commented on the issue:

    https://github.com/apache/carbondata/pull/2437
 
    Build Success with Spark 2.1.0, Please check CI http://136.243.101.176:8080/job/ApacheCarbonPRBuilder1/6776/



---

[qiuchenjian-2]

Github user manishgupta88 commented on the issue:

    https://github.com/apache/carbondata/pull/2437
 
    @ravipesala ..handled review comments. Kindly review and merge


---

[qiuchenjian-2]

Github user CarbonDataQA commented on the issue:

    https://github.com/apache/carbondata/pull/2437
 
    Build Failed  with Spark 2.1.0, Please check CI http://136.243.101.176:8080/job/ApacheCarbonPRBuilder1/6796/



---