use of com.google.cloud.bigquery.storage.v1beta2.CreateReadSessionRequest in project beam by apache.
the class BigQueryIOStorageReadTest method testTableSourceInitialSplit_EmptyTable.
@Test
public void testTableSourceInitialSplit_EmptyTable() throws Exception {
fakeDatasetService.createDataset("foo.com:project", "dataset", "", "", null);
TableReference tableRef = BigQueryHelpers.parseTableSpec("foo.com:project:dataset.table");
Table table = new Table().setTableReference(tableRef).setNumBytes(1024L * 1024L).setSchema(new TableSchema());
fakeDatasetService.createTable(table);
CreateReadSessionRequest expectedRequest = CreateReadSessionRequest.newBuilder().setParent("projects/project-id").setReadSession(ReadSession.newBuilder().setTable("projects/foo.com:project/datasets/dataset/tables/table")).setMaxStreamCount(1024).build();
ReadSession emptyReadSession = ReadSession.newBuilder().build();
StorageClient fakeStorageClient = mock(StorageClient.class);
when(fakeStorageClient.createReadSession(expectedRequest)).thenReturn(emptyReadSession);
BigQueryStorageTableSource<TableRow> tableSource = BigQueryStorageTableSource.create(ValueProvider.StaticValueProvider.of(tableRef), null, null, new TableRowParser(), TableRowJsonCoder.of(), new FakeBigQueryServices().withDatasetService(fakeDatasetService).withStorageClient(fakeStorageClient));
List<? extends BoundedSource<TableRow>> sources = tableSource.split(1024L, options);
assertTrue(sources.isEmpty());
}
use of com.google.cloud.bigquery.storage.v1beta2.CreateReadSessionRequest in project beam by apache.
the class BigQueryIOStorageQueryTest method testQuerySourceInitialSplit_NoReferencedTables.
/**
* This test simulates the scenario where the SQL text which is executed by the query job doesn't
* by itself refer to any tables (e.g. "SELECT 17 AS value"), and thus there are no referenced
* tables when the dry run of the query is performed.
*/
@Test
public void testQuerySourceInitialSplit_NoReferencedTables() throws Exception {
Table queryResultTable = new Table().setSchema(new TableSchema().setFields(ImmutableList.of(new TableFieldSchema().setName("name").setType("STRING"), new TableFieldSchema().setName("number").setType("INTEGER")))).setNumBytes(1024L * 1024L);
String encodedQuery = FakeBigQueryServices.encodeQueryResult(queryResultTable);
fakeJobService.expectDryRunQuery(options.getProject(), encodedQuery, new JobStatistics().setQuery(new JobStatistics2().setTotalBytesProcessed(1024L * 1024L).setReferencedTables(ImmutableList.of())));
String stepUuid = "testStepUuid";
TableReference tempTableReference = createTempTableReference(options.getProject(), BigQueryResourceNaming.createJobIdPrefix(options.getJobName(), stepUuid, JobType.QUERY), Optional.empty());
CreateReadSessionRequest expectedRequest = CreateReadSessionRequest.newBuilder().setParent("projects/" + options.getProject()).setReadSession(ReadSession.newBuilder().setTable(BigQueryHelpers.toTableResourceName(tempTableReference))).setMaxStreamCount(1024).build();
Schema sessionSchema = SchemaBuilder.record("__root__").fields().name("name").type().nullable().stringType().noDefault().name("number").type().nullable().longType().noDefault().endRecord();
ReadSession.Builder builder = ReadSession.newBuilder().setAvroSchema(AvroSchema.newBuilder().setSchema(sessionSchema.toString())).setDataFormat(DataFormat.AVRO);
for (int i = 0; i < 1024; i++) {
builder.addStreams(ReadStream.newBuilder().setName("stream-" + i));
}
StorageClient fakeStorageClient = mock(StorageClient.class);
when(fakeStorageClient.createReadSession(expectedRequest)).thenReturn(builder.build());
BigQueryStorageQuerySource<TableRow> querySource = BigQueryStorageQuerySource.create(stepUuid, ValueProvider.StaticValueProvider.of(encodedQuery), /* flattenResults = */
true, /* useLegacySql = */
true, /* priority = */
QueryPriority.BATCH, /* location = */
null, /* queryTempDataset = */
null, /* kmsKey = */
null, null, new TableRowParser(), TableRowJsonCoder.of(), new FakeBigQueryServices().withDatasetService(fakeDatasetService).withJobService(fakeJobService).withStorageClient(fakeStorageClient));
List<? extends BoundedSource<TableRow>> sources = querySource.split(1024, options);
assertEquals(1024, sources.size());
}
use of com.google.cloud.bigquery.storage.v1beta2.CreateReadSessionRequest in project beam by apache.
the class BigQueryIOStorageReadTest method testReadFromBigQueryIOArrow.
@Test
public void testReadFromBigQueryIOArrow() throws Exception {
fakeDatasetService.createDataset("foo.com:project", "dataset", "", "", null);
TableReference tableRef = BigQueryHelpers.parseTableSpec("foo.com:project:dataset.table");
Table table = new Table().setTableReference(tableRef).setNumBytes(10L).setSchema(TABLE_SCHEMA);
fakeDatasetService.createTable(table);
CreateReadSessionRequest expectedCreateReadSessionRequest = CreateReadSessionRequest.newBuilder().setParent("projects/project-id").setReadSession(ReadSession.newBuilder().setTable("projects/foo.com:project/datasets/dataset/tables/table").setDataFormat(DataFormat.ARROW)).setMaxStreamCount(10).build();
ReadSession readSession = ReadSession.newBuilder().setName("readSessionName").setArrowSchema(ArrowSchema.newBuilder().setSerializedSchema(serializeArrowSchema(ARROW_SCHEMA)).build()).addStreams(ReadStream.newBuilder().setName("streamName")).setDataFormat(DataFormat.ARROW).build();
ReadRowsRequest expectedReadRowsRequest = ReadRowsRequest.newBuilder().setReadStream("streamName").build();
List<String> names = Arrays.asList("A", "B", "C", "D");
List<Long> values = Arrays.asList(1L, 2L, 3L, 4L);
List<ReadRowsResponse> readRowsResponses = Lists.newArrayList(createResponseArrow(ARROW_SCHEMA, names.subList(0, 2), values.subList(0, 2), 0.0, 0.50), createResponseArrow(ARROW_SCHEMA, names.subList(2, 4), values.subList(2, 4), 0.5, 0.75));
StorageClient fakeStorageClient = mock(StorageClient.class, withSettings().serializable());
when(fakeStorageClient.createReadSession(expectedCreateReadSessionRequest)).thenReturn(readSession);
when(fakeStorageClient.readRows(expectedReadRowsRequest, "")).thenReturn(new FakeBigQueryServerStream<>(readRowsResponses));
PCollection<KV<String, Long>> output = p.apply(BigQueryIO.read(new ParseKeyValue()).from("foo.com:project:dataset.table").withMethod(Method.DIRECT_READ).withFormat(DataFormat.ARROW).withTestServices(new FakeBigQueryServices().withDatasetService(fakeDatasetService).withStorageClient(fakeStorageClient)));
PAssert.that(output).containsInAnyOrder(ImmutableList.of(KV.of("A", 1L), KV.of("B", 2L), KV.of("C", 3L), KV.of("D", 4L)));
p.run();
}
use of com.google.cloud.bigquery.storage.v1beta2.CreateReadSessionRequest in project beam by apache.
the class BigQueryIOStorageReadTest method doTableSourceInitialSplitTest.
private void doTableSourceInitialSplitTest(long bundleSize, int streamCount) throws Exception {
fakeDatasetService.createDataset("foo.com:project", "dataset", "", "", null);
TableReference tableRef = BigQueryHelpers.parseTableSpec("foo.com:project:dataset.table");
Table table = new Table().setTableReference(tableRef).setNumBytes(1024L * 1024L).setSchema(TABLE_SCHEMA);
fakeDatasetService.createTable(table);
CreateReadSessionRequest expectedRequest = CreateReadSessionRequest.newBuilder().setParent("projects/project-id").setReadSession(ReadSession.newBuilder().setTable("projects/foo.com:project/datasets/dataset/tables/table")).setMaxStreamCount(streamCount).build();
ReadSession.Builder builder = ReadSession.newBuilder().setAvroSchema(AvroSchema.newBuilder().setSchema(AVRO_SCHEMA_STRING)).setDataFormat(DataFormat.AVRO);
for (int i = 0; i < streamCount; i++) {
builder.addStreams(ReadStream.newBuilder().setName("stream-" + i));
}
StorageClient fakeStorageClient = mock(StorageClient.class);
when(fakeStorageClient.createReadSession(expectedRequest)).thenReturn(builder.build());
BigQueryStorageTableSource<TableRow> tableSource = BigQueryStorageTableSource.create(ValueProvider.StaticValueProvider.of(tableRef), null, null, new TableRowParser(), TableRowJsonCoder.of(), new FakeBigQueryServices().withDatasetService(fakeDatasetService).withStorageClient(fakeStorageClient));
List<? extends BoundedSource<TableRow>> sources = tableSource.split(bundleSize, options);
assertEquals(streamCount, sources.size());
}
use of com.google.cloud.bigquery.storage.v1beta2.CreateReadSessionRequest in project beam by apache.
the class BigQueryIOStorageReadTest method testTableSourceInitialSplit_WithSelectedFieldsAndRowRestriction.
@Test
public void testTableSourceInitialSplit_WithSelectedFieldsAndRowRestriction() throws Exception {
fakeDatasetService.createDataset("foo.com:project", "dataset", "", "", null);
TableReference tableRef = BigQueryHelpers.parseTableSpec("foo.com:project:dataset.table");
Table table = new Table().setTableReference(tableRef).setNumBytes(100L).setSchema(TABLE_SCHEMA);
fakeDatasetService.createTable(table);
CreateReadSessionRequest expectedRequest = CreateReadSessionRequest.newBuilder().setParent("projects/project-id").setReadSession(ReadSession.newBuilder().setTable("projects/foo.com:project/datasets/dataset/tables/table").setReadOptions(ReadSession.TableReadOptions.newBuilder().addSelectedFields("name").setRowRestriction("number > 5"))).setMaxStreamCount(10).build();
ReadSession.Builder builder = ReadSession.newBuilder().setAvroSchema(AvroSchema.newBuilder().setSchema(TRIMMED_AVRO_SCHEMA_STRING)).setDataFormat(DataFormat.AVRO);
for (int i = 0; i < 10; i++) {
builder.addStreams(ReadStream.newBuilder().setName("stream-" + i));
}
StorageClient fakeStorageClient = mock(StorageClient.class);
when(fakeStorageClient.createReadSession(expectedRequest)).thenReturn(builder.build());
BigQueryStorageTableSource<TableRow> tableSource = BigQueryStorageTableSource.create(ValueProvider.StaticValueProvider.of(tableRef), StaticValueProvider.of(Lists.newArrayList("name")), StaticValueProvider.of("number > 5"), new TableRowParser(), TableRowJsonCoder.of(), new FakeBigQueryServices().withDatasetService(fakeDatasetService).withStorageClient(fakeStorageClient));
List<? extends BoundedSource<TableRow>> sources = tableSource.split(10L, options);
assertEquals(10L, sources.size());
}
Aggregations