Examples with OrcFileTail com.facebook.presto.orc.metadata.OrcFileTail used on opensource projects

Example 1 with OrcFileTail

use of com.facebook.presto.orc.metadata.OrcFileTail in project presto by prestodb.

the class TestStorageOrcFileTailSource method testReadDwrfStripeCacheIfEnabledButAbsent.

@Test
public void testReadDwrfStripeCacheIfEnabledButAbsent() throws IOException {
    FileOutputStream out = new FileOutputStream(file.getFile());
    // write the footer and post script
    DwrfProto.Footer.Builder footer = DwrfProto.Footer.newBuilder();
    DwrfProto.PostScript.Builder postScript = DwrfProto.PostScript.newBuilder().setCompression(NONE);
    writeTail(footer, postScript, out);
    out.close();
    // read the file tail with the enabled "read dwrf stripe cache" feature
    StorageOrcFileTailSource src = new StorageOrcFileTailSource(FOOTER_READ_SIZE_IN_BYTES, true);
    OrcDataSource orcDataSource = createFileOrcDataSource();
    OrcFileTail orcFileTail = src.getOrcFileTail(orcDataSource, metadataReader, Optional.empty(), false);
    assertEquals(orcFileTail.getMetadataSize(), 0);
    DwrfProto.Footer actualFooter = readFooter(orcFileTail);
    assertEquals(actualFooter, footer.build());
    // the feature is enabled, but file doesn't have the stripe cache
    assertFalse(orcFileTail.getDwrfStripeCacheData().isPresent());
}

Example 2 with OrcFileTail

use of com.facebook.presto.orc.metadata.OrcFileTail in project presto by prestodb.

the class StorageModule method createOrcFileTailSource.

@Singleton
@Provides
public OrcFileTailSource createOrcFileTailSource(OrcCacheConfig orcCacheConfig, MBeanExporter exporter) {
    int expectedFileTailSizeInBytes = toIntExact(orcCacheConfig.getExpectedFileTailSize().toBytes());
    boolean dwrfStripeCacheEnabled = orcCacheConfig.isDwrfStripeCacheEnabled();
    OrcFileTailSource orcFileTailSource = new StorageOrcFileTailSource(expectedFileTailSizeInBytes, dwrfStripeCacheEnabled);
    if (orcCacheConfig.isFileTailCacheEnabled()) {
        Cache<OrcDataSourceId, OrcFileTail> cache = CacheBuilder.newBuilder().maximumWeight(orcCacheConfig.getFileTailCacheSize().toBytes()).weigher((id, tail) -> ((OrcFileTail) tail).getTotalSize()).expireAfterAccess(orcCacheConfig.getFileTailCacheTtlSinceLastAccess().toMillis(), TimeUnit.MILLISECONDS).recordStats().build();
        CacheStatsMBean cacheStatsMBean = new CacheStatsMBean(cache);
        orcFileTailSource = new CachingOrcFileTailSource(orcFileTailSource, cache);
        exporter.export(generatedNameOf(CacheStatsMBean.class, connectorId + "_OrcFileTail"), cacheStatsMBean);
    }
    return orcFileTailSource;
}

Example 3 with OrcFileTail

use of com.facebook.presto.orc.metadata.OrcFileTail in project presto by prestodb.

the class AbstractTestOrcReader method testCaching.

@Test
public void testCaching() throws Exception {
    Cache<OrcDataSourceId, OrcFileTail> orcFileTailCache = CacheBuilder.newBuilder().maximumWeight(new DataSize(1, MEGABYTE).toBytes()).weigher((id, tail) -> ((OrcFileTail) tail).getFooterSize() + ((OrcFileTail) tail).getMetadataSize()).expireAfterAccess(new Duration(10, MINUTES).toMillis(), MILLISECONDS).recordStats().build();
    OrcFileTailSource orcFileTailSource = new CachingOrcFileTailSource(new StorageOrcFileTailSource(), orcFileTailCache);
    Cache<StripeId, Slice> stripeFootercache = CacheBuilder.newBuilder().maximumWeight(new DataSize(1, MEGABYTE).toBytes()).weigher((id, footer) -> ((Slice) footer).length()).expireAfterAccess(new Duration(10, MINUTES).toMillis(), MILLISECONDS).recordStats().build();
    Cache<StripeStreamId, Slice> stripeStreamCache = CacheBuilder.newBuilder().maximumWeight(new DataSize(1, MEGABYTE).toBytes()).weigher((id, stream) -> ((Slice) stream).length()).expireAfterAccess(new Duration(10, MINUTES).toMillis(), MILLISECONDS).recordStats().build();
    Optional<Cache<StripeStreamId, List<RowGroupIndex>>> rowGroupIndexCache = Optional.of(CacheBuilder.newBuilder().maximumWeight(new DataSize(1, MEGABYTE).toBytes()).weigher((id, rowGroupIndices) -> toIntExact(((List<RowGroupIndex>) rowGroupIndices).stream().mapToLong(RowGroupIndex::getRetainedSizeInBytes).sum())).expireAfterAccess(new Duration(10, MINUTES).toMillis(), MILLISECONDS).recordStats().build());
    StripeMetadataSource stripeMetadataSource = new CachingStripeMetadataSource(new StorageStripeMetadataSource(), stripeFootercache, stripeStreamCache, rowGroupIndexCache);
    try (TempFile tempFile = createTempFile(10001)) {
        OrcBatchRecordReader storageReader = createCustomOrcRecordReader(tempFile, ORC, OrcPredicate.TRUE, ImmutableList.of(BIGINT), INITIAL_BATCH_SIZE, orcFileTailSource, stripeMetadataSource, true, ImmutableMap.of(), false);
        assertEquals(orcFileTailCache.stats().missCount(), 1);
        assertEquals(orcFileTailCache.stats().hitCount(), 0);
        OrcBatchRecordReader cacheReader = createCustomOrcRecordReader(tempFile, ORC, OrcPredicate.TRUE, ImmutableList.of(BIGINT), INITIAL_BATCH_SIZE, orcFileTailSource, stripeMetadataSource, true, ImmutableMap.of(), false);
        assertEquals(orcFileTailCache.stats().missCount(), 1);
        assertEquals(orcFileTailCache.stats().hitCount(), 1);
        assertEquals(storageReader.getRetainedSizeInBytes(), cacheReader.getRetainedSizeInBytes());
        assertEquals(storageReader.getFileRowCount(), cacheReader.getFileRowCount());
        assertEquals(storageReader.getSplitLength(), cacheReader.getSplitLength());
        storageReader.nextBatch();
        assertEquals(stripeFootercache.stats().missCount(), 1);
        assertEquals(stripeFootercache.stats().hitCount(), 0);
        assertEquals(stripeStreamCache.stats().missCount(), 2);
        assertEquals(stripeStreamCache.stats().hitCount(), 0);
        assertEquals(rowGroupIndexCache.get().stats().missCount(), 1);
        assertEquals(rowGroupIndexCache.get().stats().hitCount(), 0);
        cacheReader.nextBatch();
        assertEquals(stripeFootercache.stats().missCount(), 1);
        assertEquals(stripeFootercache.stats().hitCount(), 1);
        assertEquals(stripeStreamCache.stats().missCount(), 2);
        assertEquals(stripeStreamCache.stats().hitCount(), 2);
        assertEquals(rowGroupIndexCache.get().stats().missCount(), 1);
        assertEquals(rowGroupIndexCache.get().stats().hitCount(), 1);
        assertEquals(storageReader.readBlock(0).getInt(0), cacheReader.readBlock(0).getInt(0));
    }
}

Example 4 with OrcFileTail

use of com.facebook.presto.orc.metadata.OrcFileTail in project presto by prestodb.

the class HiveClientModule method createOrcFileTailSource.

@Singleton
@Provides
public OrcFileTailSource createOrcFileTailSource(OrcCacheConfig orcCacheConfig, MBeanExporter exporter) {
    int expectedFileTailSizeInBytes = toIntExact(orcCacheConfig.getExpectedFileTailSize().toBytes());
    boolean dwrfStripeCacheEnabled = orcCacheConfig.isDwrfStripeCacheEnabled();
    OrcFileTailSource orcFileTailSource = new StorageOrcFileTailSource(expectedFileTailSizeInBytes, dwrfStripeCacheEnabled);
    if (orcCacheConfig.isFileTailCacheEnabled()) {
        Cache<OrcDataSourceId, OrcFileTail> cache = CacheBuilder.newBuilder().maximumWeight(orcCacheConfig.getFileTailCacheSize().toBytes()).weigher((id, tail) -> ((OrcFileTail) tail).getFooterSize() + ((OrcFileTail) tail).getMetadataSize()).expireAfterAccess(orcCacheConfig.getFileTailCacheTtlSinceLastAccess().toMillis(), MILLISECONDS).recordStats().build();
        CacheStatsMBean cacheStatsMBean = new CacheStatsMBean(cache);
        orcFileTailSource = new CachingOrcFileTailSource(orcFileTailSource, cache);
        exporter.export(generatedNameOf(CacheStatsMBean.class, connectorId + "_OrcFileTail"), cacheStatsMBean);
    }
    return orcFileTailSource;
}

Example 5 with OrcFileTail

use of com.facebook.presto.orc.metadata.OrcFileTail in project presto by prestodb.

the class IcebergModule method createOrcFileTailSource.

@Singleton
@Provides
public OrcFileTailSource createOrcFileTailSource(OrcCacheConfig orcCacheConfig, MBeanExporter exporter) {
    int expectedFileTailSizeInBytes = toIntExact(orcCacheConfig.getExpectedFileTailSize().toBytes());
    boolean dwrfStripeCacheEnabled = orcCacheConfig.isDwrfStripeCacheEnabled();
    OrcFileTailSource orcFileTailSource = new StorageOrcFileTailSource(expectedFileTailSizeInBytes, dwrfStripeCacheEnabled);
    if (orcCacheConfig.isFileTailCacheEnabled()) {
        Cache<OrcDataSourceId, OrcFileTail> cache = CacheBuilder.newBuilder().maximumWeight(orcCacheConfig.getFileTailCacheSize().toBytes()).weigher((id, tail) -> ((OrcFileTail) tail).getFooterSize() + ((OrcFileTail) tail).getMetadataSize()).expireAfterAccess(orcCacheConfig.getFileTailCacheTtlSinceLastAccess().toMillis(), MILLISECONDS).recordStats().build();
        CacheStatsMBean cacheStatsMBean = new CacheStatsMBean(cache);
        orcFileTailSource = new CachingOrcFileTailSource(orcFileTailSource, cache);
        exporter.export(generatedNameOf(CacheStatsMBean.class, connectorId + "_OrcFileTail"), cacheStatsMBean);
    }
    return orcFileTailSource;
}