Examples with RowGroupMetadata org.apache.drill.exec.store.parquet.metadata.MetadataBase.RowGroupMetadata used on opensource projects

Example 1 with RowGroupMetadata

use of org.apache.drill.exec.store.parquet.metadata.MetadataBase.RowGroupMetadata in project drill by apache.

the class ParquetReaderUtility method transformBinaryInMetadataCache.

/**
 * Transforms values for min / max binary statistics to byte array.
 * Transformation logic depends on metadata file version.
 *
 * @param parquetTableMetadata table metadata that should be corrected
 * @param readerConfig parquet reader config
 */
public static void transformBinaryInMetadataCache(ParquetTableMetadataBase parquetTableMetadata, ParquetReaderConfig readerConfig) {
    // Looking for the names of the columns with BINARY data type
    // in the metadata cache file for V2 and all v3 versions
    Set<List<String>> columnsNames = getBinaryColumnsNames(parquetTableMetadata);
    boolean allowBinaryMetadata = allowBinaryMetadata(parquetTableMetadata.getDrillVersion(), readerConfig);
    // Setting Min / Max values for ParquetTableMetadata_v1
    MetadataVersion metadataVersion = new MetadataVersion(parquetTableMetadata.getMetadataVersion());
    if (metadataVersion.isEqualTo(1, 0)) {
        for (ParquetFileMetadata file : parquetTableMetadata.getFiles()) {
            for (RowGroupMetadata rowGroupMetadata : file.getRowGroups()) {
                Long rowCount = rowGroupMetadata.getRowCount();
                for (ColumnMetadata columnMetadata : rowGroupMetadata.getColumns()) {
                    if (columnMetadata.getPrimitiveType() == PrimitiveTypeName.BINARY || columnMetadata.getPrimitiveType() == PrimitiveTypeName.FIXED_LEN_BYTE_ARRAY) {
                        setMinMaxValues(columnMetadata, rowCount, allowBinaryMetadata, false);
                    }
                }
            }
        }
        return;
    }
    // Variables needed for debugging only
    Stopwatch timer = logger.isDebugEnabled() ? Stopwatch.createStarted() : null;
    int maxRowGroups = 0;
    int minRowGroups = Integer.MAX_VALUE;
    int maxNumColumns = 0;
    // Setting Min / Max values for V2, V3 and V4 versions; for versions V3_3 and above need to do decoding
    boolean needDecoding = metadataVersion.isAtLeast(3, 3);
    for (ParquetFileMetadata file : parquetTableMetadata.getFiles()) {
        if (timer != null) {
            // for debugging only
            maxRowGroups = Math.max(maxRowGroups, file.getRowGroups().size());
            minRowGroups = Math.min(minRowGroups, file.getRowGroups().size());
        }
        for (RowGroupMetadata rowGroupMetadata : file.getRowGroups()) {
            Long rowCount = rowGroupMetadata.getRowCount();
            if (timer != null) {
                // for debugging only
                maxNumColumns = Math.max(maxNumColumns, rowGroupMetadata.getColumns().size());
            }
            for (ColumnMetadata columnMetadata : rowGroupMetadata.getColumns()) {
                if (columnsNames.contains(Arrays.asList(columnMetadata.getName()))) {
                    setMinMaxValues(columnMetadata, rowCount, allowBinaryMetadata, needDecoding);
                }
            }
        }
    }
    if (timer != null) {
        // log a debug message and stop the timer
        String reportRG = 1 == maxRowGroups ? "1 rowgroup" : "between " + minRowGroups + "-" + maxRowGroups + "rowgroups";
        logger.debug("Transforming binary in metadata cache took {} ms ({} files, {} per file, max {} columns)", timer.elapsed(TimeUnit.MILLISECONDS), parquetTableMetadata.getFiles().size(), reportRG, maxNumColumns);
        timer.stop();
    }
}

Example 2 with RowGroupMetadata

use of org.apache.drill.exec.store.parquet.metadata.MetadataBase.RowGroupMetadata in project drill by apache.

the class ParquetReaderUtility method correctDatesInMetadataCache.

public static void correctDatesInMetadataCache(ParquetTableMetadataBase parquetTableMetadata) {
    MetadataVersion metadataVersion = new MetadataVersion(parquetTableMetadata.getMetadataVersion());
    DateCorruptionStatus cacheFileCanContainsCorruptDates = metadataVersion.isAtLeast(3, 0) ? DateCorruptionStatus.META_SHOWS_NO_CORRUPTION : DateCorruptionStatus.META_UNCLEAR_TEST_VALUES;
    if (cacheFileCanContainsCorruptDates == DateCorruptionStatus.META_UNCLEAR_TEST_VALUES) {
        // Looking for the DATE data type of column names in the metadata cache file ("metadata_version" : "v2")
        String[] names = new String[0];
        if (metadataVersion.isEqualTo(2, 0)) {
            for (ColumnTypeMetadata_v2 columnTypeMetadata : ((ParquetTableMetadata_v2) parquetTableMetadata).columnTypeInfo.values()) {
                if (OriginalType.DATE.equals(columnTypeMetadata.originalType)) {
                    names = columnTypeMetadata.name;
                }
            }
        }
        for (ParquetFileMetadata file : parquetTableMetadata.getFiles()) {
            // Drill has only ever written a single row group per file, only need to correct the statistics
            // on the first row group
            RowGroupMetadata rowGroupMetadata = file.getRowGroups().get(0);
            Long rowCount = rowGroupMetadata.getRowCount();
            for (ColumnMetadata columnMetadata : rowGroupMetadata.getColumns()) {
                // Setting Min/Max values for ParquetTableMetadata_v1
                if (metadataVersion.isEqualTo(1, 0)) {
                    OriginalType originalType = columnMetadata.getOriginalType();
                    if (OriginalType.DATE.equals(originalType) && columnMetadata.hasSingleValue(rowCount) && (Integer) columnMetadata.getMaxValue() > ParquetReaderUtility.DATE_CORRUPTION_THRESHOLD) {
                        int newMinMax = ParquetReaderUtility.autoCorrectCorruptedDate((Integer) columnMetadata.getMaxValue());
                        columnMetadata.setMax(newMinMax);
                        columnMetadata.setMin(newMinMax);
                    }
                } else // Setting Max values for ParquetTableMetadata_v2
                if (metadataVersion.isEqualTo(2, 0) && columnMetadata.getName() != null && Arrays.equals(columnMetadata.getName(), names) && columnMetadata.hasSingleValue(rowCount) && (Integer) columnMetadata.getMaxValue() > ParquetReaderUtility.DATE_CORRUPTION_THRESHOLD) {
                    int newMax = ParquetReaderUtility.autoCorrectCorruptedDate((Integer) columnMetadata.getMaxValue());
                    columnMetadata.setMax(newMax);
                }
            }
        }
    }
}