Search in sources :

Example 31 with InputRow

use of io.druid.data.input.InputRow in project druid by druid-io.

the class TimeseriesBenchmark method setup.

@Setup
public void setup() throws IOException {
    log.info("SETUP CALLED AT " + System.currentTimeMillis());
    if (ComplexMetrics.getSerdeForType("hyperUnique") == null) {
        ComplexMetrics.registerSerde("hyperUnique", new HyperUniquesSerde(HyperLogLogHash.getDefault()));
    }
    executorService = Execs.multiThreaded(numSegments, "TimeseriesThreadPool");
    setupQueries();
    String[] schemaQuery = schemaAndQuery.split("\\.");
    String schemaName = schemaQuery[0];
    String queryName = schemaQuery[1];
    schemaInfo = BenchmarkSchemas.SCHEMA_MAP.get(schemaName);
    query = SCHEMA_QUERY_MAP.get(schemaName).get(queryName);
    incIndexes = new ArrayList<>();
    for (int i = 0; i < numSegments; i++) {
        log.info("Generating rows for segment " + i);
        BenchmarkDataGenerator gen = new BenchmarkDataGenerator(schemaInfo.getColumnSchemas(), RNG_SEED + i, schemaInfo.getDataInterval(), rowsPerSegment);
        IncrementalIndex incIndex = makeIncIndex();
        for (int j = 0; j < rowsPerSegment; j++) {
            InputRow row = gen.nextRow();
            if (j % 10000 == 0) {
                log.info(j + " rows generated.");
            }
            incIndex.add(row);
        }
        log.info(rowsPerSegment + " rows generated");
        incIndexes.add(incIndex);
    }
    tmpDir = Files.createTempDir();
    log.info("Using temp dir: " + tmpDir.getAbsolutePath());
    qIndexes = new ArrayList<>();
    for (int i = 0; i < numSegments; i++) {
        File indexFile = INDEX_MERGER_V9.persist(incIndexes.get(i), tmpDir, new IndexSpec());
        QueryableIndex qIndex = INDEX_IO.loadIndex(indexFile);
        qIndexes.add(qIndex);
    }
    factory = new TimeseriesQueryRunnerFactory(new TimeseriesQueryQueryToolChest(QueryBenchmarkUtil.NoopIntervalChunkingQueryRunnerDecorator()), new TimeseriesQueryEngine(), QueryBenchmarkUtil.NOOP_QUERYWATCHER);
}
Also used : IndexSpec(io.druid.segment.IndexSpec) IncrementalIndex(io.druid.segment.incremental.IncrementalIndex) OnheapIncrementalIndex(io.druid.segment.incremental.OnheapIncrementalIndex) BenchmarkDataGenerator(io.druid.benchmark.datagen.BenchmarkDataGenerator) HyperUniquesSerde(io.druid.query.aggregation.hyperloglog.HyperUniquesSerde) TimeseriesQueryQueryToolChest(io.druid.query.timeseries.TimeseriesQueryQueryToolChest) TimeseriesQueryEngine(io.druid.query.timeseries.TimeseriesQueryEngine) TimeseriesQueryRunnerFactory(io.druid.query.timeseries.TimeseriesQueryRunnerFactory) QueryableIndex(io.druid.segment.QueryableIndex) InputRow(io.druid.data.input.InputRow) File(java.io.File) Setup(org.openjdk.jmh.annotations.Setup)

Example 32 with InputRow

use of io.druid.data.input.InputRow in project druid by druid-io.

the class SearchBenchmark method setup.

@Setup
public void setup() throws IOException {
    log.info("SETUP CALLED AT " + +System.currentTimeMillis());
    if (ComplexMetrics.getSerdeForType("hyperUnique") == null) {
        ComplexMetrics.registerSerde("hyperUnique", new HyperUniquesSerde(HyperLogLogHash.getDefault()));
    }
    executorService = Execs.multiThreaded(numSegments, "SearchThreadPool");
    setupQueries();
    String[] schemaQuery = schemaAndQuery.split("\\.");
    String schemaName = schemaQuery[0];
    String queryName = schemaQuery[1];
    schemaInfo = BenchmarkSchemas.SCHEMA_MAP.get(schemaName);
    queryBuilder = SCHEMA_QUERY_MAP.get(schemaName).get(queryName);
    queryBuilder.limit(limit);
    query = queryBuilder.build();
    incIndexes = new ArrayList<>();
    for (int i = 0; i < numSegments; i++) {
        log.info("Generating rows for segment " + i);
        BenchmarkDataGenerator gen = new BenchmarkDataGenerator(schemaInfo.getColumnSchemas(), System.currentTimeMillis(), schemaInfo.getDataInterval(), rowsPerSegment);
        IncrementalIndex incIndex = makeIncIndex();
        for (int j = 0; j < rowsPerSegment; j++) {
            InputRow row = gen.nextRow();
            if (j % 10000 == 0) {
                log.info(j + " rows generated.");
            }
            incIndex.add(row);
        }
        incIndexes.add(incIndex);
    }
    tmpDir = Files.createTempDir();
    log.info("Using temp dir: " + tmpDir.getAbsolutePath());
    qIndexes = new ArrayList<>();
    for (int i = 0; i < numSegments; i++) {
        File indexFile = INDEX_MERGER_V9.persist(incIndexes.get(i), tmpDir, new IndexSpec());
        QueryableIndex qIndex = INDEX_IO.loadIndex(indexFile);
        qIndexes.add(qIndex);
    }
    final SearchQueryConfig config = new SearchQueryConfig().withOverrides(query);
    factory = new SearchQueryRunnerFactory(new SearchStrategySelector(Suppliers.ofInstance(config)), new SearchQueryQueryToolChest(config, QueryBenchmarkUtil.NoopIntervalChunkingQueryRunnerDecorator()), QueryBenchmarkUtil.NOOP_QUERYWATCHER);
}
Also used : IndexSpec(io.druid.segment.IndexSpec) SearchQueryConfig(io.druid.query.search.search.SearchQueryConfig) IncrementalIndex(io.druid.segment.incremental.IncrementalIndex) OnheapIncrementalIndex(io.druid.segment.incremental.OnheapIncrementalIndex) BenchmarkDataGenerator(io.druid.benchmark.datagen.BenchmarkDataGenerator) HyperUniquesSerde(io.druid.query.aggregation.hyperloglog.HyperUniquesSerde) SearchQueryRunnerFactory(io.druid.query.search.SearchQueryRunnerFactory) SearchStrategySelector(io.druid.query.search.SearchStrategySelector) SearchQueryQueryToolChest(io.druid.query.search.SearchQueryQueryToolChest) QueryableIndex(io.druid.segment.QueryableIndex) InputRow(io.druid.data.input.InputRow) File(java.io.File) Setup(org.openjdk.jmh.annotations.Setup)

Example 33 with InputRow

use of io.druid.data.input.InputRow in project druid by druid-io.

the class TopNBenchmark method setup.

@Setup
public void setup() throws IOException {
    log.info("SETUP CALLED AT " + System.currentTimeMillis());
    if (ComplexMetrics.getSerdeForType("hyperUnique") == null) {
        ComplexMetrics.registerSerde("hyperUnique", new HyperUniquesSerde(HyperLogLogHash.getDefault()));
    }
    executorService = Execs.multiThreaded(numSegments, "TopNThreadPool");
    setupQueries();
    String[] schemaQuery = schemaAndQuery.split("\\.");
    String schemaName = schemaQuery[0];
    String queryName = schemaQuery[1];
    schemaInfo = BenchmarkSchemas.SCHEMA_MAP.get(schemaName);
    queryBuilder = SCHEMA_QUERY_MAP.get(schemaName).get(queryName);
    queryBuilder.threshold(threshold);
    query = queryBuilder.build();
    incIndexes = new ArrayList<>();
    for (int i = 0; i < numSegments; i++) {
        log.info("Generating rows for segment " + i);
        BenchmarkDataGenerator gen = new BenchmarkDataGenerator(schemaInfo.getColumnSchemas(), RNG_SEED + i, schemaInfo.getDataInterval(), rowsPerSegment);
        IncrementalIndex incIndex = makeIncIndex();
        for (int j = 0; j < rowsPerSegment; j++) {
            InputRow row = gen.nextRow();
            if (j % 10000 == 0) {
                log.info(j + " rows generated.");
            }
            incIndex.add(row);
        }
        incIndexes.add(incIndex);
    }
    tmpDir = Files.createTempDir();
    log.info("Using temp dir: " + tmpDir.getAbsolutePath());
    qIndexes = new ArrayList<>();
    for (int i = 0; i < numSegments; i++) {
        File indexFile = INDEX_MERGER_V9.persist(incIndexes.get(i), tmpDir, new IndexSpec());
        QueryableIndex qIndex = INDEX_IO.loadIndex(indexFile);
        qIndexes.add(qIndex);
    }
    factory = new TopNQueryRunnerFactory(new StupidPool<>("TopNBenchmark-compute-bufferPool", new OffheapBufferGenerator("compute", 250000000), 0, Integer.MAX_VALUE), new TopNQueryQueryToolChest(new TopNQueryConfig(), QueryBenchmarkUtil.NoopIntervalChunkingQueryRunnerDecorator()), QueryBenchmarkUtil.NOOP_QUERYWATCHER);
}
Also used : IndexSpec(io.druid.segment.IndexSpec) IncrementalIndex(io.druid.segment.incremental.IncrementalIndex) OnheapIncrementalIndex(io.druid.segment.incremental.OnheapIncrementalIndex) BenchmarkDataGenerator(io.druid.benchmark.datagen.BenchmarkDataGenerator) HyperUniquesSerde(io.druid.query.aggregation.hyperloglog.HyperUniquesSerde) OffheapBufferGenerator(io.druid.offheap.OffheapBufferGenerator) TopNQueryConfig(io.druid.query.topn.TopNQueryConfig) QueryableIndex(io.druid.segment.QueryableIndex) InputRow(io.druid.data.input.InputRow) TopNQueryRunnerFactory(io.druid.query.topn.TopNQueryRunnerFactory) StupidPool(io.druid.collections.StupidPool) TopNQueryQueryToolChest(io.druid.query.topn.TopNQueryQueryToolChest) File(java.io.File) Setup(org.openjdk.jmh.annotations.Setup)

Example 34 with InputRow

use of io.druid.data.input.InputRow in project druid by druid-io.

the class BenchmarkDataGeneratorTest method testDiscreteUniform.

@Test
public void testDiscreteUniform() throws Exception {
    List<BenchmarkColumnSchema> schemas = new ArrayList<>();
    RowValueTracker tracker = new RowValueTracker();
    schemas.add(BenchmarkColumnSchema.makeDiscreteUniform("dimA", ValueType.STRING, false, 1, null, 10, 20));
    schemas.add(BenchmarkColumnSchema.makeEnumeratedDiscreteUniform("dimB", ValueType.STRING, false, 4, null, Arrays.<Object>asList("Hello", "World", "Foo", "Bar")));
    schemas.add(BenchmarkColumnSchema.makeDiscreteUniform("dimC", ValueType.STRING, false, 1, 0.50, 10, 20));
    schemas.add(BenchmarkColumnSchema.makeDiscreteUniform("dimD", ValueType.FLOAT, false, 1, null, 100, 120));
    BenchmarkDataGenerator dataGenerator = new BenchmarkDataGenerator(schemas, 9999, 0, 0, 1000.0);
    for (int i = 0; i < 100; i++) {
        InputRow row = dataGenerator.nextRow();
        //System.out.println("U-ROW: " + row);
        tracker.addRow(row);
    }
    tracker.printStuff();
}
Also used : ArrayList(java.util.ArrayList) BenchmarkDataGenerator(io.druid.benchmark.datagen.BenchmarkDataGenerator) InputRow(io.druid.data.input.InputRow) BenchmarkColumnSchema(io.druid.benchmark.datagen.BenchmarkColumnSchema) Test(org.junit.Test)

Example 35 with InputRow

use of io.druid.data.input.InputRow in project druid by druid-io.

the class BenchmarkDataGeneratorTest method testSequential.

@Test
public void testSequential() throws Exception {
    List<BenchmarkColumnSchema> schemas = new ArrayList<>();
    RowValueTracker tracker = new RowValueTracker();
    schemas.add(BenchmarkColumnSchema.makeSequential("dimA", ValueType.STRING, false, 1, null, 10, 20));
    schemas.add(BenchmarkColumnSchema.makeEnumeratedSequential("dimB", ValueType.STRING, false, 1, null, Arrays.<Object>asList("Hello", "World", "Foo", "Bar")));
    schemas.add(BenchmarkColumnSchema.makeSequential("dimC", ValueType.STRING, false, 1, 0.50, 30, 40));
    BenchmarkDataGenerator dataGenerator = new BenchmarkDataGenerator(schemas, 9999, 0, 0, 1000.0);
    for (int i = 0; i < 100; i++) {
        InputRow row = dataGenerator.nextRow();
        //System.out.println("S-ROW: " + row);
        tracker.addRow(row);
    }
    tracker.printStuff();
}
Also used : ArrayList(java.util.ArrayList) BenchmarkDataGenerator(io.druid.benchmark.datagen.BenchmarkDataGenerator) InputRow(io.druid.data.input.InputRow) BenchmarkColumnSchema(io.druid.benchmark.datagen.BenchmarkColumnSchema) Test(org.junit.Test)

Aggregations

InputRow (io.druid.data.input.InputRow)81 Test (org.junit.Test)35 MapBasedInputRow (io.druid.data.input.MapBasedInputRow)24 BenchmarkDataGenerator (io.druid.benchmark.datagen.BenchmarkDataGenerator)22 File (java.io.File)18 Setup (org.openjdk.jmh.annotations.Setup)15 HyperUniquesSerde (io.druid.query.aggregation.hyperloglog.HyperUniquesSerde)14 Firehose (io.druid.data.input.Firehose)12 OnheapIncrementalIndex (io.druid.segment.incremental.OnheapIncrementalIndex)12 IndexSpec (io.druid.segment.IndexSpec)11 ArrayList (java.util.ArrayList)11 IncrementalIndex (io.druid.segment.incremental.IncrementalIndex)10 DateTime (org.joda.time.DateTime)10 QueryableIndex (io.druid.segment.QueryableIndex)9 IOException (java.io.IOException)9 BenchmarkColumnSchema (io.druid.benchmark.datagen.BenchmarkColumnSchema)8 Interval (org.joda.time.Interval)8 ParseException (io.druid.java.util.common.parsers.ParseException)7 AggregatorFactory (io.druid.query.aggregation.AggregatorFactory)6 DataSegment (io.druid.timeline.DataSegment)5