use of org.apache.carbondata.core.datastore.chunk.impl.FixedLengthDimensionDataChunk in project carbondata by apache.
the class CompressedDimensionChunkFileBasedReaderV3 method convertToDimensionChunk.
/**
* Below method will be used to convert the compressed dimension chunk raw data to actual data
*
* @param dimensionRawColumnChunk dimension raw chunk
* @param pageNumber number
* @return DimensionColumnDataChunk
*/
@Override
public DimensionColumnDataChunk convertToDimensionChunk(DimensionRawColumnChunk dimensionRawColumnChunk, int pageNumber) throws IOException {
byte[] dataPage = null;
int[] invertedIndexes = null;
int[] invertedIndexesReverse = null;
int[] rlePage = null;
// data chunk of page
DataChunk2 dimensionColumnChunk = null;
// data chunk of blocklet column
DataChunk3 dataChunk3 = dimensionRawColumnChunk.getDataChunkV3();
// get the data buffer
ByteBuffer rawData = dimensionRawColumnChunk.getRawData();
dimensionColumnChunk = dataChunk3.getData_chunk_list().get(pageNumber);
// calculating the start point of data
// as buffer can contain multiple column data, start point will be datachunkoffset +
// data chunk length + page offset
int copySourcePoint = dimensionRawColumnChunk.getOffSet() + dimensionChunksLength.get(dimensionRawColumnChunk.getBlockletId()) + dataChunk3.getPage_offset().get(pageNumber);
// first read the data and uncompressed it
dataPage = COMPRESSOR.unCompressByte(rawData.array(), copySourcePoint, dimensionColumnChunk.data_page_length);
copySourcePoint += dimensionColumnChunk.data_page_length;
// if row id block is present then read the row id chunk and uncompress it
if (hasEncoding(dimensionColumnChunk.encoders, Encoding.INVERTED_INDEX)) {
invertedIndexes = CarbonUtil.getUnCompressColumnIndex(dimensionColumnChunk.rowid_page_length, rawData, copySourcePoint);
copySourcePoint += dimensionColumnChunk.rowid_page_length;
// get the reverse index
invertedIndexesReverse = getInvertedReverseIndex(invertedIndexes);
}
//then actual data based on rle block
if (hasEncoding(dimensionColumnChunk.encoders, Encoding.RLE)) {
rlePage = CarbonUtil.getIntArray(rawData, copySourcePoint, dimensionColumnChunk.rle_page_length);
// uncompress the data with rle indexes
dataPage = UnBlockIndexer.uncompressData(dataPage, rlePage, eachColumnValueSize[dimensionRawColumnChunk.getBlockletId()]);
rlePage = null;
}
// fill chunk attributes
DimensionColumnDataChunk columnDataChunk = null;
if (dimensionColumnChunk.isRowMajor()) {
// to store fixed length column chunk values
columnDataChunk = new ColumnGroupDimensionDataChunk(dataPage, eachColumnValueSize[dimensionRawColumnChunk.getBlockletId()], dimensionRawColumnChunk.getRowCount()[pageNumber]);
} else // and set to data chunk instance
if (!hasEncoding(dimensionColumnChunk.encoders, Encoding.DICTIONARY)) {
columnDataChunk = new VariableLengthDimensionDataChunk(dataPage, invertedIndexes, invertedIndexesReverse, dimensionRawColumnChunk.getRowCount()[pageNumber]);
} else {
// to store fixed length column chunk values
columnDataChunk = new FixedLengthDimensionDataChunk(dataPage, invertedIndexes, invertedIndexesReverse, dimensionRawColumnChunk.getRowCount()[pageNumber], eachColumnValueSize[dimensionRawColumnChunk.getBlockletId()]);
}
return columnDataChunk;
}
use of org.apache.carbondata.core.datastore.chunk.impl.FixedLengthDimensionDataChunk in project carbondata by apache.
the class CarbonUtilTest method testToGetNextLesserValueToTarget.
@Test
public void testToGetNextLesserValueToTarget() {
byte[] dataChunks = { 7, 7, 7, 8, 9 };
byte[] compareValues = { 7 };
FixedLengthDimensionDataChunk fixedLengthDataChunk = new FixedLengthDimensionDataChunk(dataChunks, null, null, 5, 1);
int result = CarbonUtil.nextLesserValueToTarget(2, fixedLengthDataChunk, compareValues);
assertEquals(result, -1);
}
use of org.apache.carbondata.core.datastore.chunk.impl.FixedLengthDimensionDataChunk in project carbondata by apache.
the class CarbonUtilTest method testToGetNextLesserValue.
@Test
public void testToGetNextLesserValue() {
byte[] dataChunks = { 5, 6, 7, 8, 9 };
byte[] compareValues = { 7 };
FixedLengthDimensionDataChunk fixedLengthDataChunk = new FixedLengthDimensionDataChunk(dataChunks, null, null, 5, 1);
int result = CarbonUtil.nextLesserValueToTarget(2, fixedLengthDataChunk, compareValues);
assertEquals(result, 1);
}
use of org.apache.carbondata.core.datastore.chunk.impl.FixedLengthDimensionDataChunk in project carbondata by apache.
the class CarbonUtilTest method testToGetFirstIndexUsingBinarySearchWithCompareToLessThan0.
@Test
public void testToGetFirstIndexUsingBinarySearchWithCompareToLessThan0() {
byte[] dataChunks = { 10, 20, 30, 40, 50, 60 };
byte[] compareValue = { 30 };
FixedLengthDimensionDataChunk fixedLengthDimensionDataChunk = new FixedLengthDimensionDataChunk(dataChunks, null, null, 6, 1);
int result = CarbonUtil.getFirstIndexUsingBinarySearch(fixedLengthDimensionDataChunk, 1, 3, compareValue, false);
assertEquals(2, result);
}
use of org.apache.carbondata.core.datastore.chunk.impl.FixedLengthDimensionDataChunk in project carbondata by apache.
the class CarbonUtilTest method testToGetFirstIndexUsingBinarySearchWithMatchUpLimitTrue.
@Test
public void testToGetFirstIndexUsingBinarySearchWithMatchUpLimitTrue() {
byte[] dataChunks = { 10, 10, 10, 40, 50, 60 };
byte[] compareValue = { 10 };
FixedLengthDimensionDataChunk fixedLengthDimensionDataChunk = new FixedLengthDimensionDataChunk(dataChunks, null, null, 6, 1);
int result = CarbonUtil.getFirstIndexUsingBinarySearch(fixedLengthDimensionDataChunk, 1, 3, compareValue, true);
assertEquals(2, result);
}
Aggregations