use of org.apache.carbondata.format.BlockletBTreeIndex in project carbondata by apache.
the class CarbonMetadataUtil method getBlockletIndex.
public static BlockletIndex getBlockletIndex(List<NodeHolder> nodeHolderList, List<CarbonMeasure> carbonMeasureList) {
BlockletMinMaxIndex blockletMinMaxIndex = new BlockletMinMaxIndex();
// Calculating min/max for every each column.
byte[][] minCol = nodeHolderList.get(0).getColumnMinData().clone();
byte[][] maxCol = nodeHolderList.get(0).getColumnMaxData().clone();
for (NodeHolder nodeHolder : nodeHolderList) {
byte[][] columnMaxData = nodeHolder.getColumnMaxData();
byte[][] columnMinData = nodeHolder.getColumnMinData();
for (int i = 0; i < maxCol.length; i++) {
if (ByteUtil.UnsafeComparer.INSTANCE.compareTo(columnMaxData[i], maxCol[i]) > 0) {
maxCol[i] = columnMaxData[i];
}
if (ByteUtil.UnsafeComparer.INSTANCE.compareTo(columnMinData[i], minCol[i]) < 0) {
minCol[i] = columnMinData[i];
}
}
}
// Writing min/max to thrift file
for (byte[] max : maxCol) {
blockletMinMaxIndex.addToMax_values(ByteBuffer.wrap(max));
}
for (byte[] min : minCol) {
blockletMinMaxIndex.addToMin_values(ByteBuffer.wrap(min));
}
byte[][] measureMaxValue = nodeHolderList.get(0).getMeasureColumnMaxData().clone();
byte[][] measureMinValue = nodeHolderList.get(0).getMeasureColumnMinData().clone();
byte[] minVal = null;
byte[] maxVal = null;
for (int i = 1; i < nodeHolderList.size(); i++) {
for (int j = 0; j < measureMinValue.length; j++) {
minVal = nodeHolderList.get(i).getMeasureColumnMinData()[j];
maxVal = nodeHolderList.get(i).getMeasureColumnMaxData()[j];
if (compareMeasureData(measureMaxValue[j], maxVal, carbonMeasureList.get(j).getDataType()) < 0) {
measureMaxValue[j] = maxVal.clone();
}
if (compareMeasureData(measureMinValue[j], minVal, carbonMeasureList.get(j).getDataType()) > 0) {
measureMinValue[j] = minVal.clone();
}
}
}
for (byte[] max : measureMaxValue) {
blockletMinMaxIndex.addToMax_values(ByteBuffer.wrap(max));
}
for (byte[] min : measureMinValue) {
blockletMinMaxIndex.addToMin_values(ByteBuffer.wrap(min));
}
BlockletBTreeIndex blockletBTreeIndex = new BlockletBTreeIndex();
blockletBTreeIndex.setStart_key(nodeHolderList.get(0).getStartKey());
blockletBTreeIndex.setEnd_key(nodeHolderList.get(nodeHolderList.size() - 1).getEndKey());
BlockletIndex blockletIndex = new BlockletIndex();
blockletIndex.setMin_max_index(blockletMinMaxIndex);
blockletIndex.setB_tree_index(blockletBTreeIndex);
return blockletIndex;
}
use of org.apache.carbondata.format.BlockletBTreeIndex in project carbondata by apache.
the class DataFileFooterConverterTest method testReadDataFileFooter.
@Test
public void testReadDataFileFooter() throws Exception {
DataFileFooterConverter dataFileFooterConverter = new DataFileFooterConverter();
DataFileFooter dataFileFooter = new DataFileFooter();
List<Integer> column_cardinalities = new ArrayList<>();
column_cardinalities.add(new Integer("1"));
column_cardinalities.add(new Integer("2"));
column_cardinalities.add(new Integer("3"));
org.apache.carbondata.format.SegmentInfo segmentInfo1 = new org.apache.carbondata.format.SegmentInfo(3, column_cardinalities);
List<Encoding> encoders = new ArrayList<>();
encoders.add(Encoding.INVERTED_INDEX);
encoders.add(Encoding.BIT_PACKED);
encoders.add(Encoding.DELTA);
encoders.add(Encoding.DICTIONARY);
encoders.add(Encoding.DIRECT_DICTIONARY);
encoders.add(Encoding.RLE);
ColumnSchema columnSchema = new ColumnSchema(DataType.INT, "column", "3", true, encoders, true);
ColumnSchema columnSchema1 = new ColumnSchema(DataType.ARRAY, "column", "3", true, encoders, true);
ColumnSchema columnSchema2 = new ColumnSchema(DataType.DECIMAL, "column", "3", true, encoders, true);
ColumnSchema columnSchema3 = new ColumnSchema(DataType.DOUBLE, "column", "3", true, encoders, true);
ColumnSchema columnSchema4 = new ColumnSchema(DataType.LONG, "column", "3", true, encoders, true);
ColumnSchema columnSchema5 = new ColumnSchema(DataType.SHORT, "column", "3", true, encoders, true);
ColumnSchema columnSchema6 = new ColumnSchema(DataType.STRUCT, "column", "3", true, encoders, true);
ColumnSchema columnSchema7 = new ColumnSchema(DataType.STRING, "column", "3", true, encoders, true);
final List<ColumnSchema> columnSchemas = new ArrayList<>();
columnSchemas.add(columnSchema);
columnSchemas.add(columnSchema1);
columnSchemas.add(columnSchema2);
columnSchemas.add(columnSchema3);
columnSchemas.add(columnSchema4);
columnSchemas.add(columnSchema5);
columnSchemas.add(columnSchema6);
columnSchemas.add(columnSchema7);
org.apache.carbondata.format.BlockletIndex blockletIndex1 = new org.apache.carbondata.format.BlockletIndex();
List<org.apache.carbondata.format.BlockletIndex> blockletIndexArrayList = new ArrayList<>();
blockletIndexArrayList.add(blockletIndex1);
org.apache.carbondata.format.BlockletInfo blockletInfo = new org.apache.carbondata.format.BlockletInfo();
List<org.apache.carbondata.format.BlockletInfo> blockletInfoArrayList = new ArrayList<>();
blockletInfoArrayList.add(blockletInfo);
final FileFooter fileFooter = new FileFooter(1, 3, columnSchemas, segmentInfo1, blockletIndexArrayList);
fileFooter.setBlocklet_info_list(blockletInfoArrayList);
BlockletBTreeIndex blockletBTreeIndex = new BlockletBTreeIndex();
blockletBTreeIndex.setStart_key("1".getBytes());
blockletBTreeIndex.setEnd_key("3".getBytes());
blockletIndex1.setB_tree_index(blockletBTreeIndex);
BlockletMinMaxIndex blockletMinMaxIndex = new BlockletMinMaxIndex();
blockletMinMaxIndex.setMax_values(Arrays.asList(ByteBuffer.allocate(1).put((byte) 2)));
blockletMinMaxIndex.setMin_values(Arrays.asList(ByteBuffer.allocate(1).put((byte) 1)));
blockletIndex1.setMin_max_index(blockletMinMaxIndex);
new MockUp<FileFactory>() {
@SuppressWarnings("unused")
@Mock
public FileFactory.FileType getFileType(String path) {
return FileFactory.FileType.LOCAL;
}
@SuppressWarnings("unused")
@Mock
public FileReader getFileHolder(FileFactory.FileType fileType) {
return new FileReaderImpl();
}
};
new MockUp<FileReaderImpl>() {
@SuppressWarnings("unused")
@Mock
public long readLong(String filePath, long offset) {
return 1;
}
};
new MockUp<CarbonFooterReader>() {
@SuppressWarnings("unused")
@Mock
public FileFooter readFooter() throws IOException {
return fileFooter;
}
};
SegmentInfo segmentInfo = new SegmentInfo();
int[] arr = { 1, 2, 3 };
segmentInfo.setColumnCardinality(arr);
dataFileFooter.setNumberOfRows(3);
dataFileFooter.setSegmentInfo(segmentInfo);
TableBlockInfo info = new TableBlockInfo("/file.carbondata", 1, "0", new String[0], 1, ColumnarFormatVersion.V1, null);
DataFileFooter result = dataFileFooterConverter.readDataFileFooter(info);
assertEquals(result.getNumberOfRows(), 3);
}
use of org.apache.carbondata.format.BlockletBTreeIndex in project carbondata by apache.
the class CarbonMetadataUtil method getBlockletIndex.
public static BlockletIndex getBlockletIndex(org.apache.carbondata.core.metadata.blocklet.index.BlockletIndex info) {
BlockletMinMaxIndex blockletMinMaxIndex = new BlockletMinMaxIndex();
for (int i = 0; i < info.getMinMaxIndex().getMaxValues().length; i++) {
blockletMinMaxIndex.addToMax_values(ByteBuffer.wrap(info.getMinMaxIndex().getMaxValues()[i]));
blockletMinMaxIndex.addToMin_values(ByteBuffer.wrap(info.getMinMaxIndex().getMinValues()[i]));
}
BlockletBTreeIndex blockletBTreeIndex = new BlockletBTreeIndex();
blockletBTreeIndex.setStart_key(info.getBtreeIndex().getStartKey());
blockletBTreeIndex.setEnd_key(info.getBtreeIndex().getEndKey());
BlockletIndex blockletIndex = new BlockletIndex();
blockletIndex.setMin_max_index(blockletMinMaxIndex);
blockletIndex.setB_tree_index(blockletBTreeIndex);
return blockletIndex;
}
use of org.apache.carbondata.format.BlockletBTreeIndex in project carbondata by apache.
the class CarbonMetadataUtil method getBlockletIndex.
public static BlockletIndex getBlockletIndex(List<EncodedTablePage> encodedTablePageList, List<CarbonMeasure> carbonMeasureList) {
BlockletMinMaxIndex blockletMinMaxIndex = new BlockletMinMaxIndex();
// Calculating min/max for every each column.
TablePageStatistics stats = new TablePageStatistics(encodedTablePageList.get(0).getDimensions(), encodedTablePageList.get(0).getMeasures());
byte[][] minCol = stats.getDimensionMinValue().clone();
byte[][] maxCol = stats.getDimensionMaxValue().clone();
for (EncodedTablePage encodedTablePage : encodedTablePageList) {
stats = new TablePageStatistics(encodedTablePage.getDimensions(), encodedTablePage.getMeasures());
byte[][] columnMaxData = stats.getDimensionMaxValue();
byte[][] columnMinData = stats.getDimensionMinValue();
for (int i = 0; i < maxCol.length; i++) {
if (ByteUtil.UnsafeComparer.INSTANCE.compareTo(columnMaxData[i], maxCol[i]) > 0) {
maxCol[i] = columnMaxData[i];
}
if (ByteUtil.UnsafeComparer.INSTANCE.compareTo(columnMinData[i], minCol[i]) < 0) {
minCol[i] = columnMinData[i];
}
}
}
// Writing min/max to thrift file
for (byte[] max : maxCol) {
blockletMinMaxIndex.addToMax_values(ByteBuffer.wrap(max));
}
for (byte[] min : minCol) {
blockletMinMaxIndex.addToMin_values(ByteBuffer.wrap(min));
}
stats = new TablePageStatistics(encodedTablePageList.get(0).getDimensions(), encodedTablePageList.get(0).getMeasures());
byte[][] measureMaxValue = stats.getMeasureMaxValue().clone();
byte[][] measureMinValue = stats.getMeasureMinValue().clone();
byte[] minVal = null;
byte[] maxVal = null;
for (int i = 1; i < encodedTablePageList.size(); i++) {
for (int j = 0; j < measureMinValue.length; j++) {
stats = new TablePageStatistics(encodedTablePageList.get(i).getDimensions(), encodedTablePageList.get(i).getMeasures());
minVal = stats.getMeasureMinValue()[j];
maxVal = stats.getMeasureMaxValue()[j];
if (compareMeasureData(measureMaxValue[j], maxVal, carbonMeasureList.get(j).getDataType()) < 0) {
measureMaxValue[j] = maxVal.clone();
}
if (compareMeasureData(measureMinValue[j], minVal, carbonMeasureList.get(j).getDataType()) > 0) {
measureMinValue[j] = minVal.clone();
}
}
}
for (byte[] max : measureMaxValue) {
blockletMinMaxIndex.addToMax_values(ByteBuffer.wrap(max));
}
for (byte[] min : measureMinValue) {
blockletMinMaxIndex.addToMin_values(ByteBuffer.wrap(min));
}
BlockletBTreeIndex blockletBTreeIndex = new BlockletBTreeIndex();
byte[] startKey = encodedTablePageList.get(0).getPageKey().serializeStartKey();
blockletBTreeIndex.setStart_key(startKey);
byte[] endKey = encodedTablePageList.get(encodedTablePageList.size() - 1).getPageKey().serializeEndKey();
blockletBTreeIndex.setEnd_key(endKey);
BlockletIndex blockletIndex = new BlockletIndex();
blockletIndex.setMin_max_index(blockletMinMaxIndex);
blockletIndex.setB_tree_index(blockletBTreeIndex);
return blockletIndex;
}
use of org.apache.carbondata.format.BlockletBTreeIndex in project carbondata by apache.
the class CarbonMetadataUtil method getBlockletIndex.
private static BlockletIndex getBlockletIndex(BlockletInfoColumnar info) {
BlockletMinMaxIndex blockletMinMaxIndex = new BlockletMinMaxIndex();
for (byte[] max : info.getColumnMaxData()) {
blockletMinMaxIndex.addToMax_values(ByteBuffer.wrap(max));
}
for (byte[] min : info.getColumnMinData()) {
blockletMinMaxIndex.addToMin_values(ByteBuffer.wrap(min));
}
BlockletBTreeIndex blockletBTreeIndex = new BlockletBTreeIndex();
blockletBTreeIndex.setStart_key(info.getStartKey());
blockletBTreeIndex.setEnd_key(info.getEndKey());
BlockletIndex blockletIndex = new BlockletIndex();
blockletIndex.setMin_max_index(blockletMinMaxIndex);
blockletIndex.setB_tree_index(blockletBTreeIndex);
return blockletIndex;
}
Aggregations