use of org.apache.drill.exec.rpc.UserClientConnection in project drill by axbaretto.
the class ParquetRecordReaderTest method testPerformance.
@Test
@Ignore
public void testPerformance() throws Exception {
final DrillbitContext bitContext = mock(DrillbitContext.class);
final UserClientConnection connection = mock(UserClientConnection.class);
final DrillConfig c = DrillConfig.create();
final FunctionImplementationRegistry registry = new FunctionImplementationRegistry(c);
final FragmentContextImpl context = new FragmentContextImpl(bitContext, BitControl.PlanFragment.getDefaultInstance(), connection, registry);
final String fileName = "/tmp/parquet_test_performance.parquet";
final HashMap<String, FieldInfo> fields = new HashMap<>();
final ParquetTestProperties props = new ParquetTestProperties(1, 20 * 1000 * 1000, DEFAULT_BYTES_PER_PAGE, fields);
populateFieldInfoMap(props);
final Configuration dfsConfig = new Configuration();
final List<Footer> footers = ParquetFileReader.readFooters(dfsConfig, new Path(fileName));
final Footer f = footers.iterator().next();
final List<SchemaPath> columns = Lists.newArrayList();
columns.add(new SchemaPath("_MAP.integer", ExpressionPosition.UNKNOWN));
columns.add(new SchemaPath("_MAP.bigInt", ExpressionPosition.UNKNOWN));
columns.add(new SchemaPath("_MAP.f", ExpressionPosition.UNKNOWN));
columns.add(new SchemaPath("_MAP.d", ExpressionPosition.UNKNOWN));
columns.add(new SchemaPath("_MAP.b", ExpressionPosition.UNKNOWN));
columns.add(new SchemaPath("_MAP.bin", ExpressionPosition.UNKNOWN));
columns.add(new SchemaPath("_MAP.bin2", ExpressionPosition.UNKNOWN));
int totalRowCount = 0;
final FileSystem fs = new CachedSingleFileSystem(fileName);
final BufferAllocator allocator = RootAllocatorFactory.newRoot(c);
for (int i = 0; i < 25; i++) {
final ParquetRecordReader rr = new ParquetRecordReader(context, fileName, 0, fs, CodecFactory.createDirectCodecFactory(dfsConfig, new ParquetDirectByteBufferAllocator(allocator), 0), f.getParquetMetadata(), columns, ParquetReaderUtility.DateCorruptionStatus.META_SHOWS_CORRUPTION);
final TestOutputMutator mutator = new TestOutputMutator(allocator);
rr.setup(null, mutator);
final Stopwatch watch = Stopwatch.createStarted();
int rowCount = 0;
while ((rowCount = rr.next()) > 0) {
totalRowCount += rowCount;
}
System.out.println(String.format("Time completed: %s. ", watch.elapsed(TimeUnit.MILLISECONDS)));
rr.close();
}
allocator.close();
System.out.println(String.format("Total row count %s", totalRowCount));
}
use of org.apache.drill.exec.rpc.UserClientConnection in project drill by apache.
the class TestHashJoin method testHJMockScanCommon.
private void testHJMockScanCommon(String physicalPlan, int expectedRows) throws Throwable {
final DrillbitContext bitContext = mockDrillbitContext();
final UserClientConnection connection = Mockito.mock(UserClientConnection.class);
final PhysicalPlanReader reader = PhysicalPlanReaderTestFactory.defaultPhysicalPlanReader(c);
final PhysicalPlan plan = reader.readPhysicalPlan(Files.asCharSource(DrillFileUtils.getResourceAsFile(physicalPlan), Charsets.UTF_8).read());
final FunctionImplementationRegistry registry = new FunctionImplementationRegistry(c);
final FragmentContextImpl context = new FragmentContextImpl(bitContext, PlanFragment.getDefaultInstance(), connection, registry);
final SimpleRootExec exec = new SimpleRootExec(ImplCreator.getExec(context, (FragmentRoot) plan.getSortedOperators(false).iterator().next()));
int totalRecordCount = 0;
while (exec.next()) {
totalRecordCount += exec.getRecordCount();
}
exec.close();
assertEquals(expectedRows, totalRecordCount);
if (context.getExecutorState().getFailureCause() != null) {
throw context.getExecutorState().getFailureCause();
}
assertTrue(!context.getExecutorState().isFailed());
}
use of org.apache.drill.exec.rpc.UserClientConnection in project drill by apache.
the class TestMergeJoin method simpleEqualityJoin.
@Test
// this doesn't have a sort. it also causes an infinite loop. these may or may not be related.
@Ignore
public void simpleEqualityJoin() throws Throwable {
final DrillbitContext bitContext = mockDrillbitContext();
final UserClientConnection connection = Mockito.mock(UserClientConnection.class);
final PhysicalPlanReader reader = PhysicalPlanReaderTestFactory.defaultPhysicalPlanReader(c);
final PhysicalPlan plan = reader.readPhysicalPlan(Files.asCharSource(DrillFileUtils.getResourceAsFile("/join/merge_join.json"), Charsets.UTF_8).read());
final FunctionImplementationRegistry registry = new FunctionImplementationRegistry(c);
final FragmentContextImpl context = new FragmentContextImpl(bitContext, PlanFragment.getDefaultInstance(), connection, registry);
final SimpleRootExec exec = new SimpleRootExec(ImplCreator.getExec(context, (FragmentRoot) plan.getSortedOperators(false).iterator().next()));
int totalRecordCount = 0;
final StringBuilder sb = new StringBuilder();
while (exec.next()) {
totalRecordCount += exec.getRecordCount();
for (final ValueVector v : exec) {
sb.append("[" + v.getField().getName() + "] ");
}
sb.append("\n\n");
for (int valueIdx = 0; valueIdx < exec.getRecordCount(); valueIdx++) {
final List<Object> row = new ArrayList<>();
for (final ValueVector v : exec) {
row.add(v.getAccessor().getObject(valueIdx));
}
for (final Object cell : row) {
if (cell == null) {
sb.append("<null> ");
continue;
}
final int len = cell.toString().length();
sb.append(cell);
for (int i = 0; i < (14 - len); ++i) {
sb.append(" ");
}
}
sb.append("\n");
}
sb.append("\n");
}
logger.info(sb.toString());
assertEquals(100, totalRecordCount);
if (context.getExecutorState().getFailureCause() != null) {
throw context.getExecutorState().getFailureCause();
}
assertTrue(!context.getExecutorState().isFailed());
}
use of org.apache.drill.exec.rpc.UserClientConnection in project drill by apache.
the class TestMergeJoin method testJoinBatchSize.
@Test
public void testJoinBatchSize() throws Throwable {
final DrillbitContext bitContext = mockDrillbitContext();
final UserClientConnection connection = Mockito.mock(UserClientConnection.class);
final PhysicalPlanReader reader = PhysicalPlanReaderTestFactory.defaultPhysicalPlanReader(c);
final PhysicalPlan plan = reader.readPhysicalPlan(Files.asCharSource(DrillFileUtils.getResourceAsFile("/join/join_batchsize.json"), Charsets.UTF_8).read());
final FunctionImplementationRegistry registry = new FunctionImplementationRegistry(c);
final FragmentContextImpl context = new FragmentContextImpl(bitContext, PlanFragment.getDefaultInstance(), connection, registry);
final SimpleRootExec exec = new SimpleRootExec(ImplCreator.getExec(context, (FragmentRoot) plan.getSortedOperators(false).iterator().next()));
// skip schema batch
exec.next();
while (exec.next()) {
assertEquals(100, exec.getRecordCount());
}
if (context.getExecutorState().getFailureCause() != null) {
throw context.getExecutorState().getFailureCause();
}
assertTrue(!context.getExecutorState().isFailed());
}
use of org.apache.drill.exec.rpc.UserClientConnection in project drill by apache.
the class TestMergeJoin method orderedEqualityInnerJoin.
@Test
@Ignore
public void orderedEqualityInnerJoin() throws Throwable {
final DrillbitContext bitContext = mockDrillbitContext();
final UserClientConnection connection = Mockito.mock(UserClientConnection.class);
final PhysicalPlanReader reader = PhysicalPlanReaderTestFactory.defaultPhysicalPlanReader(c, new StoragePluginRegistryImpl(bitContext));
final PhysicalPlan plan = reader.readPhysicalPlan(Files.asCharSource(DrillFileUtils.getResourceAsFile("/join/merge_inner_single_batch.json"), Charsets.UTF_8).read().replace("#{LEFT_FILE}", DrillFileUtils.getResourceAsFile("/join/merge_single_batch.left.json").toURI().toString()).replace("#{RIGHT_FILE}", DrillFileUtils.getResourceAsFile("/join/merge_single_batch.right.json").toURI().toString()));
final FunctionImplementationRegistry registry = new FunctionImplementationRegistry(c);
final FragmentContextImpl context = new FragmentContextImpl(bitContext, PlanFragment.getDefaultInstance(), connection, registry);
final SimpleRootExec exec = new SimpleRootExec(ImplCreator.getExec(context, (FragmentRoot) plan.getSortedOperators(false).iterator().next()));
int totalRecordCount = 0;
final StringBuilder sb = new StringBuilder();
while (exec.next()) {
totalRecordCount += exec.getRecordCount();
sb.append(String.format("got next with record count: %d (total: %d):\n", exec.getRecordCount(), totalRecordCount));
sb.append(" t1 t2\n");
for (int valueIdx = 0; valueIdx < exec.getRecordCount(); valueIdx++) {
final List<Object> row = Lists.newArrayList();
for (final ValueVector v : exec) {
row.add(v.getField().getName() + ":" + v.getAccessor().getObject(valueIdx));
}
for (final Object cell : row) {
if (cell == null) {
sb.append("<null> ");
continue;
}
final int len = cell.toString().length();
sb.append(cell + " ");
for (int i = 0; i < (10 - len); ++i) {
sb.append(" ");
}
}
sb.append('\n');
}
}
assertEquals(23, totalRecordCount);
if (context.getExecutorState().getFailureCause() != null) {
throw context.getExecutorState().getFailureCause();
}
assertTrue(!context.getExecutorState().isFailed());
}
Aggregations