use of org.apache.carbondata.core.metadata.blocklet.BlockletInfo in project carbondata by apache.
the class DataFileFooterConverter2 method readDataFileFooter.
/**
* Below method will be used to convert thrift file meta to wrapper file meta
*/
@Override
public DataFileFooter readDataFileFooter(TableBlockInfo tableBlockInfo) throws IOException {
DataFileFooter dataFileFooter = new DataFileFooter();
CarbonFooterReader reader = new CarbonFooterReader(tableBlockInfo.getFilePath(), tableBlockInfo.getBlockOffset());
FileFooter footer = reader.readFooter();
dataFileFooter.setVersionId(ColumnarFormatVersion.valueOf((short) footer.getVersion()));
dataFileFooter.setNumberOfRows(footer.getNum_rows());
dataFileFooter.setSegmentInfo(getSegmentInfo(footer.getSegment_info()));
List<ColumnSchema> columnSchemaList = new ArrayList<ColumnSchema>();
List<org.apache.carbondata.format.ColumnSchema> table_columns = footer.getTable_columns();
for (int i = 0; i < table_columns.size(); i++) {
columnSchemaList.add(thriftColumnSchmeaToWrapperColumnSchema(table_columns.get(i)));
}
dataFileFooter.setColumnInTable(columnSchemaList);
List<org.apache.carbondata.format.BlockletIndex> leaf_node_indices_Thrift = footer.getBlocklet_index_list();
List<BlockletIndex> blockletIndexList = new ArrayList<BlockletIndex>();
for (int i = 0; i < leaf_node_indices_Thrift.size(); i++) {
BlockletIndex blockletIndex = getBlockletIndex(leaf_node_indices_Thrift.get(i));
blockletIndexList.add(blockletIndex);
}
List<org.apache.carbondata.format.BlockletInfo2> leaf_node_infos_Thrift = footer.getBlocklet_info_list2();
List<BlockletInfo> blockletInfoList = new ArrayList<BlockletInfo>();
for (int i = 0; i < leaf_node_infos_Thrift.size(); i++) {
BlockletInfo blockletInfo = getBlockletInfo(leaf_node_infos_Thrift.get(i), getNumberOfDimensionColumns(columnSchemaList));
blockletInfo.setBlockletIndex(blockletIndexList.get(i));
blockletInfoList.add(blockletInfo);
}
dataFileFooter.setBlockletList(blockletInfoList);
dataFileFooter.setBlockletIndex(getBlockletIndexForDataFileFooter(blockletIndexList));
return dataFileFooter;
}
use of org.apache.carbondata.core.metadata.blocklet.BlockletInfo in project carbondata by apache.
the class DataFileFooterConverterV3 method readDataFileFooter.
/**
* Below method will be used to convert thrift file meta to wrapper file meta
* This method will read the footer from footer offset present in the data file
* 1. It will read the header from carbon data file, header starts from 0 offset
* 2. It will set the stream offset
* 3. It will read the footer data from file
* 4. parse the footer to thrift object
* 5. convert to wrapper object
*
* @param tableBlockInfo
* table block info
* @return data file footer
*/
@Override
public DataFileFooter readDataFileFooter(TableBlockInfo tableBlockInfo) throws IOException {
DataFileFooter dataFileFooter = new DataFileFooter();
CarbonHeaderReader carbonHeaderReader = new CarbonHeaderReader(tableBlockInfo.getFilePath());
FileHeader fileHeader = carbonHeaderReader.readHeader();
CarbonFooterReaderV3 reader = new CarbonFooterReaderV3(tableBlockInfo.getFilePath(), tableBlockInfo.getBlockOffset());
FileFooter3 footer = reader.readFooterVersion3();
dataFileFooter.setVersionId(ColumnarFormatVersion.valueOf((short) fileHeader.getVersion()));
dataFileFooter.setNumberOfRows(footer.getNum_rows());
dataFileFooter.setSegmentInfo(getSegmentInfo(footer.getSegment_info()));
dataFileFooter.setSchemaUpdatedTimeStamp(fileHeader.getTime_stamp());
List<ColumnSchema> columnSchemaList = new ArrayList<ColumnSchema>();
List<org.apache.carbondata.format.ColumnSchema> table_columns = fileHeader.getColumn_schema();
for (int i = 0; i < table_columns.size(); i++) {
columnSchemaList.add(thriftColumnSchmeaToWrapperColumnSchema(table_columns.get(i)));
}
dataFileFooter.setColumnInTable(columnSchemaList);
List<org.apache.carbondata.format.BlockletIndex> leaf_node_indices_Thrift = footer.getBlocklet_index_list();
List<BlockletIndex> blockletIndexList = new ArrayList<BlockletIndex>();
for (int i = 0; i < leaf_node_indices_Thrift.size(); i++) {
BlockletIndex blockletIndex = getBlockletIndex(leaf_node_indices_Thrift.get(i));
blockletIndexList.add(blockletIndex);
}
List<org.apache.carbondata.format.BlockletInfo3> leaf_node_infos_Thrift = footer.getBlocklet_info_list3();
List<BlockletInfo> blockletInfoList = new ArrayList<BlockletInfo>();
for (int i = 0; i < leaf_node_infos_Thrift.size(); i++) {
BlockletInfo blockletInfo = getBlockletInfo(leaf_node_infos_Thrift.get(i), CarbonUtil.getNumberOfDimensionColumns(columnSchemaList));
blockletInfo.setBlockletIndex(blockletIndexList.get(i));
blockletInfoList.add(blockletInfo);
}
dataFileFooter.setBlockletList(blockletInfoList);
dataFileFooter.setBlockletIndex(getBlockletIndexForDataFileFooter(blockletIndexList));
return dataFileFooter;
}
use of org.apache.carbondata.core.metadata.blocklet.BlockletInfo in project carbondata by apache.
the class DataFileFooterConverterV3 method getBlockletInfo.
/**
* Below method is to convert the blocklet info of the thrift to wrapper
* blocklet info
*
* @param blockletInfoThrift blocklet info of the thrift
* @return blocklet info wrapper
*/
public BlockletInfo getBlockletInfo(org.apache.carbondata.format.BlockletInfo3 blockletInfoThrift, int numberOfDimensionColumns) {
BlockletInfo blockletInfo = new BlockletInfo();
List<Long> dimensionColumnChunkOffsets = blockletInfoThrift.getColumn_data_chunks_offsets().subList(0, numberOfDimensionColumns);
List<Long> measureColumnChunksOffsets = blockletInfoThrift.getColumn_data_chunks_offsets().subList(numberOfDimensionColumns, blockletInfoThrift.getColumn_data_chunks_offsets().size());
List<Integer> dimensionColumnChunkLength = blockletInfoThrift.getColumn_data_chunks_length().subList(0, numberOfDimensionColumns);
List<Integer> measureColumnChunksLength = blockletInfoThrift.getColumn_data_chunks_length().subList(numberOfDimensionColumns, blockletInfoThrift.getColumn_data_chunks_offsets().size());
blockletInfo.setDimensionChunkOffsets(dimensionColumnChunkOffsets);
blockletInfo.setMeasureChunkOffsets(measureColumnChunksOffsets);
blockletInfo.setDimensionChunksLength(dimensionColumnChunkLength);
blockletInfo.setMeasureChunksLength(measureColumnChunksLength);
blockletInfo.setNumberOfRows(blockletInfoThrift.getNum_rows());
blockletInfo.setDimensionOffset(blockletInfoThrift.getDimension_offsets());
blockletInfo.setMeasureOffsets(blockletInfoThrift.getMeasure_offsets());
blockletInfo.setNumberOfPages(blockletInfoThrift.getNumber_number_of_pages());
return blockletInfo;
}
use of org.apache.carbondata.core.metadata.blocklet.BlockletInfo in project carbondata by apache.
the class AbstractDataFileFooterConverter method getIndexInfo.
/**
* Below method will be used to get the index info from index file
*
* @param filePath file path of the index file
* @return list of index info
* @throws IOException problem while reading the index file
*/
public List<DataFileFooter> getIndexInfo(String filePath, byte[] fileData) throws IOException {
CarbonIndexFileReader indexReader = new CarbonIndexFileReader();
List<DataFileFooter> dataFileFooters = new ArrayList<DataFileFooter>();
String parentPath = filePath.substring(0, filePath.lastIndexOf("/"));
try {
// open the reader
if (fileData != null) {
indexReader.openThriftReader(fileData);
} else {
indexReader.openThriftReader(filePath);
}
// get the index header
org.apache.carbondata.format.IndexHeader readIndexHeader = indexReader.readIndexHeader();
List<ColumnSchema> columnSchemaList = new ArrayList<ColumnSchema>();
List<org.apache.carbondata.format.ColumnSchema> table_columns = readIndexHeader.getTable_columns();
for (int i = 0; i < table_columns.size(); i++) {
columnSchemaList.add(thriftColumnSchmeaToWrapperColumnSchema(table_columns.get(i)));
}
// get the segment info
SegmentInfo segmentInfo = getSegmentInfo(readIndexHeader.getSegment_info());
BlockletIndex blockletIndex = null;
DataFileFooter dataFileFooter = null;
// read the block info from file
while (indexReader.hasNext()) {
BlockIndex readBlockIndexInfo = indexReader.readBlockIndexInfo();
blockletIndex = getBlockletIndex(readBlockIndexInfo.getBlock_index());
dataFileFooter = new DataFileFooter();
TableBlockInfo tableBlockInfo = getTableBlockInfo(readBlockIndexInfo, readIndexHeader, parentPath);
dataFileFooter.setBlockletIndex(blockletIndex);
dataFileFooter.setColumnInTable(columnSchemaList);
dataFileFooter.setNumberOfRows(readBlockIndexInfo.getNum_rows());
dataFileFooter.setBlockInfo(new BlockInfo(tableBlockInfo));
dataFileFooter.setSegmentInfo(segmentInfo);
dataFileFooter.setVersionId(tableBlockInfo.getVersion());
// In case of old schema time stamp will not be found in the index header
if (readIndexHeader.isSetSchema_time_stamp()) {
dataFileFooter.setSchemaUpdatedTimeStamp(readIndexHeader.getSchema_time_stamp());
}
if (readBlockIndexInfo.isSetBlocklet_info()) {
List<BlockletInfo> blockletInfoList = new ArrayList<BlockletInfo>();
BlockletInfo blockletInfo = new DataFileFooterConverterV3().getBlockletInfo(readBlockIndexInfo.getBlocklet_info(), CarbonUtil.getNumberOfDimensionColumns(columnSchemaList));
blockletInfo.setBlockletIndex(blockletIndex);
blockletInfoList.add(blockletInfo);
dataFileFooter.setBlockletList(blockletInfoList);
}
dataFileFooters.add(dataFileFooter);
}
} finally {
indexReader.closeThriftReader();
}
return dataFileFooters;
}
use of org.apache.carbondata.core.metadata.blocklet.BlockletInfo in project carbondata by apache.
the class BlockletDetailInfo method readFields.
@Override
public void readFields(DataInput in) throws IOException {
rowCount = in.readInt();
pagesCount = in.readShort();
versionNumber = in.readShort();
blockletId = in.readShort();
dimLens = new int[in.readShort()];
for (int i = 0; i < dimLens.length; i++) {
dimLens[i] = in.readInt();
}
schemaUpdatedTimeStamp = in.readLong();
if (in.readBoolean()) {
blockletInfo = new BlockletInfo();
blockletInfo.readFields(in);
}
blockFooterOffset = in.readLong();
int bytesSize = in.readInt();
byte[] schemaArray = new byte[bytesSize];
in.readFully(schemaArray);
readColumnSchema(schemaArray);
blockSize = in.readLong();
}
Aggregations