Examples with PipelineOptions org.apache.beam.sdk.options.PipelineOptions used on opensource projects

Example 1 with PipelineOptions

use of org.apache.beam.sdk.options.PipelineOptions in project beam by apache.

the class TestCountingSourceTest method testCanResumeWithExpandedCount.

@Test
public void testCanResumeWithExpandedCount() throws IOException {
    TestCountingSource source = new TestCountingSource(1);
    PipelineOptions options = PipelineOptionsFactory.create();
    TestCountingSource.CountingSourceReader reader = source.createReader(options, null);
    assertTrue(reader.start());
    assertEquals(0L, (long) reader.getCurrent().getValue());
    assertFalse(reader.advance());
    TestCountingSource.CounterMark checkpoint = reader.getCheckpointMark();
    checkpoint.finalizeCheckpoint();
    source = new TestCountingSource(2);
    reader = source.createReader(options, checkpoint);
    assertTrue(reader.start());
    assertEquals(1L, (long) reader.getCurrent().getValue());
    assertFalse(reader.advance());
}

Example 2 with PipelineOptions

use of org.apache.beam.sdk.options.PipelineOptions in project beam by apache.

the class BigQueryIOTest method testBigQueryQuerySourceInitSplit.

@Test
public void testBigQueryQuerySourceInitSplit() throws Exception {
    TableReference dryRunTable = new TableReference();
    Job queryJob = new Job();
    JobStatistics queryJobStats = new JobStatistics();
    JobStatistics2 queryStats = new JobStatistics2();
    queryStats.setReferencedTables(ImmutableList.of(dryRunTable));
    queryJobStats.setQuery(queryStats);
    queryJob.setStatus(new JobStatus()).setStatistics(queryJobStats);
    Job extractJob = new Job();
    JobStatistics extractJobStats = new JobStatistics();
    JobStatistics4 extractStats = new JobStatistics4();
    extractStats.setDestinationUriFileCounts(ImmutableList.of(1L));
    extractJobStats.setExtract(extractStats);
    extractJob.setStatus(new JobStatus()).setStatistics(extractJobStats);
    FakeJobService fakeJobService = new FakeJobService();
    FakeDatasetService fakeDatasetService = new FakeDatasetService();
    FakeBigQueryServices fakeBqServices = new FakeBigQueryServices().withJobService(fakeJobService).withDatasetService(fakeDatasetService);
    List<TableRow> expected = ImmutableList.of(new TableRow().set("name", "a").set("number", 1L), new TableRow().set("name", "b").set("number", 2L), new TableRow().set("name", "c").set("number", 3L), new TableRow().set("name", "d").set("number", 4L), new TableRow().set("name", "e").set("number", 5L), new TableRow().set("name", "f").set("number", 6L));
    PipelineOptions options = PipelineOptionsFactory.create();
    BigQueryOptions bqOptions = options.as(BigQueryOptions.class);
    bqOptions.setProject("project");
    String stepUuid = "testStepUuid";
    TableReference tempTableReference = createTempTableReference(bqOptions.getProject(), createJobIdToken(bqOptions.getJobName(), stepUuid));
    fakeDatasetService.createDataset(bqOptions.getProject(), tempTableReference.getDatasetId(), "", "");
    fakeDatasetService.createTable(new Table().setTableReference(tempTableReference).setSchema(new TableSchema().setFields(ImmutableList.of(new TableFieldSchema().setName("name").setType("STRING"), new TableFieldSchema().setName("number").setType("INTEGER")))));
    Path baseDir = Files.createTempDirectory(tempFolder, "testBigQueryQuerySourceInitSplit");
    String query = FakeBigQueryServices.encodeQuery(expected);
    BoundedSource<TableRow> bqSource = BigQueryQuerySource.create(stepUuid, StaticValueProvider.of(query), true, /* flattenResults */
    true, /* useLegacySql */
    fakeBqServices);
    options.setTempLocation(baseDir.toString());
    TableReference queryTable = new TableReference().setProjectId(bqOptions.getProject()).setDatasetId(tempTableReference.getDatasetId()).setTableId(tempTableReference.getTableId());
    fakeJobService.expectDryRunQuery(bqOptions.getProject(), query, new JobStatistics().setQuery(new JobStatistics2().setTotalBytesProcessed(100L).setReferencedTables(ImmutableList.of(queryTable))));
    List<TableRow> read = SourceTestUtils.readFromSource(bqSource, options);
    assertThat(read, containsInAnyOrder(Iterables.toArray(expected, TableRow.class)));
    SourceTestUtils.assertSplitAtFractionBehavior(bqSource, 2, 0.3, ExpectedSplitOutcome.MUST_BE_CONSISTENT_IF_SUCCEEDS, options);
    List<? extends BoundedSource<TableRow>> sources = bqSource.split(100, options);
    assertEquals(2, sources.size());
    BoundedSource<TableRow> actual = sources.get(0);
    assertThat(actual, CoreMatchers.instanceOf(TransformingSource.class));
}

Example 3 with PipelineOptions

use of org.apache.beam.sdk.options.PipelineOptions in project beam by apache.

the class BigQueryIOTest method testBigQueryTableSourceInitSplit.

@Test
public void testBigQueryTableSourceInitSplit() throws Exception {
    FakeDatasetService fakeDatasetService = new FakeDatasetService();
    FakeJobService fakeJobService = new FakeJobService();
    FakeBigQueryServices fakeBqServices = new FakeBigQueryServices().withJobService(fakeJobService).withDatasetService(fakeDatasetService);
    List<TableRow> expected = ImmutableList.of(new TableRow().set("name", "a").set("number", 1L), new TableRow().set("name", "b").set("number", 2L), new TableRow().set("name", "c").set("number", 3L), new TableRow().set("name", "d").set("number", 4L), new TableRow().set("name", "e").set("number", 5L), new TableRow().set("name", "f").set("number", 6L));
    TableReference table = BigQueryHelpers.parseTableSpec("project:data_set.table_name");
    fakeDatasetService.createDataset("project", "data_set", "", "");
    fakeDatasetService.createTable(new Table().setTableReference(table).setSchema(new TableSchema().setFields(ImmutableList.of(new TableFieldSchema().setName("name").setType("STRING"), new TableFieldSchema().setName("number").setType("INTEGER")))));
    fakeDatasetService.insertAll(table, expected, null);
    Path baseDir = Files.createTempDirectory(tempFolder, "testBigQueryTableSourceInitSplit");
    String stepUuid = "testStepUuid";
    BoundedSource<TableRow> bqSource = BigQueryTableSource.create(stepUuid, StaticValueProvider.of(table), fakeBqServices);
    PipelineOptions options = PipelineOptionsFactory.create();
    options.setTempLocation(baseDir.toString());
    BigQueryOptions bqOptions = options.as(BigQueryOptions.class);
    bqOptions.setProject("project");
    List<TableRow> read = SourceTestUtils.readFromSource(bqSource, options);
    assertThat(read, containsInAnyOrder(Iterables.toArray(expected, TableRow.class)));
    SourceTestUtils.assertSplitAtFractionBehavior(bqSource, 2, 0.3, ExpectedSplitOutcome.MUST_BE_CONSISTENT_IF_SUCCEEDS, options);
    List<? extends BoundedSource<TableRow>> sources = bqSource.split(100, options);
    assertEquals(2, sources.size());
    // Simulate a repeated call to split(), like a Dataflow worker will sometimes do.
    sources = bqSource.split(200, options);
    assertEquals(2, sources.size());
    BoundedSource<TableRow> actual = sources.get(0);
    assertThat(actual, CoreMatchers.instanceOf(TransformingSource.class));
    // A repeated call to split() should not have caused a duplicate extract job.
    assertEquals(1, fakeJobService.getNumExtractJobCalls());
}

Example 4 with PipelineOptions

use of org.apache.beam.sdk.options.PipelineOptions in project beam by apache.

the class BigQueryIOTest method testBigQueryNoTableQuerySourceInitSplit.

@Test
public void testBigQueryNoTableQuerySourceInitSplit() throws Exception {
    TableReference dryRunTable = new TableReference();
    Job queryJob = new Job();
    JobStatistics queryJobStats = new JobStatistics();
    JobStatistics2 queryStats = new JobStatistics2();
    queryStats.setReferencedTables(ImmutableList.of(dryRunTable));
    queryJobStats.setQuery(queryStats);
    queryJob.setStatus(new JobStatus()).setStatistics(queryJobStats);
    Job extractJob = new Job();
    JobStatistics extractJobStats = new JobStatistics();
    JobStatistics4 extractStats = new JobStatistics4();
    extractStats.setDestinationUriFileCounts(ImmutableList.of(1L));
    extractJobStats.setExtract(extractStats);
    extractJob.setStatus(new JobStatus()).setStatistics(extractJobStats);
    FakeDatasetService datasetService = new FakeDatasetService();
    FakeJobService jobService = new FakeJobService();
    FakeBigQueryServices fakeBqServices = new FakeBigQueryServices().withJobService(jobService).withDatasetService(datasetService);
    PipelineOptions options = PipelineOptionsFactory.create();
    BigQueryOptions bqOptions = options.as(BigQueryOptions.class);
    bqOptions.setProject("project");
    String stepUuid = "testStepUuid";
    TableReference tempTableReference = createTempTableReference(bqOptions.getProject(), createJobIdToken(bqOptions.getJobName(), stepUuid));
    List<TableRow> expected = ImmutableList.of(new TableRow().set("name", "a").set("number", 1L), new TableRow().set("name", "b").set("number", 2L), new TableRow().set("name", "c").set("number", 3L), new TableRow().set("name", "d").set("number", 4L), new TableRow().set("name", "e").set("number", 5L), new TableRow().set("name", "f").set("number", 6L));
    datasetService.createDataset(tempTableReference.getProjectId(), tempTableReference.getDatasetId(), "", "");
    Table table = new Table().setTableReference(tempTableReference).setSchema(new TableSchema().setFields(ImmutableList.of(new TableFieldSchema().setName("name").setType("STRING"), new TableFieldSchema().setName("number").setType("INTEGER"))));
    datasetService.createTable(table);
    String query = FakeBigQueryServices.encodeQuery(expected);
    jobService.expectDryRunQuery("project", query, new JobStatistics().setQuery(new JobStatistics2().setTotalBytesProcessed(100L).setReferencedTables(ImmutableList.of(table.getTableReference()))));
    Path baseDir = Files.createTempDirectory(tempFolder, "testBigQueryNoTableQuerySourceInitSplit");
    BoundedSource<TableRow> bqSource = BigQueryQuerySource.create(stepUuid, StaticValueProvider.of(query), true, /* flattenResults */
    true, /* useLegacySql */
    fakeBqServices);
    options.setTempLocation(baseDir.toString());
    List<TableRow> read = convertBigDecimaslToLong(SourceTestUtils.readFromSource(bqSource, options));
    assertThat(read, containsInAnyOrder(Iterables.toArray(expected, TableRow.class)));
    SourceTestUtils.assertSplitAtFractionBehavior(bqSource, 2, 0.3, ExpectedSplitOutcome.MUST_BE_CONSISTENT_IF_SUCCEEDS, options);
    List<? extends BoundedSource<TableRow>> sources = bqSource.split(100, options);
    assertEquals(2, sources.size());
    BoundedSource<TableRow> actual = sources.get(0);
    assertThat(actual, CoreMatchers.instanceOf(TransformingSource.class));
}

Example 5 with PipelineOptions

use of org.apache.beam.sdk.options.PipelineOptions in project beam by apache.

the class BigQueryIOTest method testTransformingSource.

@Test
public void testTransformingSource() throws Exception {
    int numElements = 10000;
    @SuppressWarnings("deprecation") BoundedSource<Long> longSource = CountingSource.upTo(numElements);
    SerializableFunction<Long, String> toStringFn = new SerializableFunction<Long, String>() {

        @Override
        public String apply(Long input) {
            return input.toString();
        }
    };
    BoundedSource<String> stringSource = new TransformingSource<>(longSource, toStringFn, StringUtf8Coder.of());
    List<String> expected = Lists.newArrayList();
    for (int i = 0; i < numElements; i++) {
        expected.add(String.valueOf(i));
    }
    PipelineOptions options = PipelineOptionsFactory.create();
    Assert.assertThat(SourceTestUtils.readFromSource(stringSource, options), CoreMatchers.is(expected));
    SourceTestUtils.assertSplitAtFractionBehavior(stringSource, 100, 0.3, ExpectedSplitOutcome.MUST_SUCCEED_AND_BE_CONSISTENT, options);
    SourceTestUtils.assertSourcesEqualReferenceSource(stringSource, stringSource.split(100, options), options);
}