All Downloads are FREE. Search and download functionalities are using the official Maven repository.

com.palantir.atlasdb.blob.generated.DataMetadataCleanupTask Maven / Gradle / Ivy

There is a newer version: 0.1152.0
Show newest version
package com.palantir.atlasdb.blob.generated;

import java.util.Iterator;
import java.util.Map;
import java.util.HashSet;
import java.util.Set;
import java.util.stream.Collectors;

import com.google.common.collect.Sets;
import com.palantir.atlasdb.cleaner.api.OnCleanupTask;
import com.palantir.atlasdb.encoding.PtBytes;
import com.palantir.atlasdb.keyvalue.api.BatchColumnRangeSelection;
import com.palantir.atlasdb.keyvalue.api.Cell;
import com.palantir.atlasdb.keyvalue.api.Namespace;
import com.palantir.atlasdb.protos.generated.StreamPersistence.Status;
import com.palantir.atlasdb.protos.generated.StreamPersistence.StreamMetadata;
import com.palantir.atlasdb.transaction.api.Transaction;

public class DataMetadataCleanupTask implements OnCleanupTask {

    private final BlobSchemaTableFactory tables;

    public DataMetadataCleanupTask(Namespace namespace) {
        tables = BlobSchemaTableFactory.of(namespace);
    }

    @Override
    public boolean cellsCleanedUp(Transaction t, Set cells) {
        DataStreamMetadataTable metaTable = tables.getDataStreamMetadataTable(t);
        Set rows = Sets.newHashSetWithExpectedSize(cells.size());
        for (Cell cell : cells) {
            rows.add(DataStreamMetadataTable.DataStreamMetadataRow.BYTES_HYDRATOR.hydrateFromBytes(cell.getRowName()));
        }
        DataStreamIdxTable indexTable = tables.getDataStreamIdxTable(t);
        Set rowsWithNoIndexEntries =
                        getUnreferencedStreamsByIterator(indexTable, rows);
        Set toDelete = new HashSet<>();
        Map currentMetadata =
                metaTable.getMetadatas(rows);
        for (Map.Entry e : currentMetadata.entrySet()) {
            if (e.getValue().getStatus() != Status.STORED || rowsWithNoIndexEntries.contains(e.getKey())) {
                toDelete.add(e.getKey().getId());
            }
        }
        DataStreamStore.of(tables).deleteStreams(t, toDelete);
        return false;
    }

    private static Set getUnreferencedStreamsByIterator(DataStreamIdxTable indexTable, Set metadataRows) {
        Set indexRows = metadataRows.stream()
                .map(DataStreamMetadataTable.DataStreamMetadataRow::getId)
                .map(DataStreamIdxTable.DataStreamIdxRow::of)
                .collect(Collectors.toSet());
        Map> referenceIteratorByStream
                = indexTable.getRowsColumnRangeIterator(indexRows,
                        BatchColumnRangeSelection.create(PtBytes.EMPTY_BYTE_ARRAY, PtBytes.EMPTY_BYTE_ARRAY, 1));
        return referenceIteratorByStream.entrySet().stream()
                .filter(entry -> !entry.getValue().hasNext())
                .map(Map.Entry::getKey)
                .map(DataStreamIdxTable.DataStreamIdxRow::getId)
                .map(DataStreamMetadataTable.DataStreamMetadataRow::of)
                .collect(Collectors.toSet());
    }
}




© 2015 - 2024 Weber Informatics LLC | Privacy Policy