use of com.google.cloud.bigquery.storage.v1beta2.ReadSession in project beam by apache.
the class BigQueryIOStorageReadTest method testReadFromStreamSource.
@Test
public void testReadFromStreamSource() throws Exception {
ReadSession readSession = ReadSession.newBuilder().setName("readSession").setAvroSchema(AvroSchema.newBuilder().setSchema(AVRO_SCHEMA_STRING)).build();
ReadRowsRequest expectedRequest = ReadRowsRequest.newBuilder().setReadStream("readStream").build();
List<GenericRecord> records = Lists.newArrayList(createRecord("A", 1, AVRO_SCHEMA), createRecord("B", 2, AVRO_SCHEMA), createRecord("C", 3, AVRO_SCHEMA));
List<ReadRowsResponse> responses = Lists.newArrayList(createResponse(AVRO_SCHEMA, records.subList(0, 2), 0.0, 0.50), createResponse(AVRO_SCHEMA, records.subList(2, 3), 0.5, 0.75));
StorageClient fakeStorageClient = mock(StorageClient.class);
when(fakeStorageClient.readRows(expectedRequest, "")).thenReturn(new FakeBigQueryServerStream<>(responses));
BigQueryStorageStreamSource<TableRow> streamSource = BigQueryStorageStreamSource.create(readSession, ReadStream.newBuilder().setName("readStream").build(), TABLE_SCHEMA, new TableRowParser(), TableRowJsonCoder.of(), new FakeBigQueryServices().withStorageClient(fakeStorageClient));
List<TableRow> rows = new ArrayList<>();
BoundedReader<TableRow> reader = streamSource.createReader(options);
for (boolean hasNext = reader.start(); hasNext; hasNext = reader.advance()) {
rows.add(reader.getCurrent());
}
System.out.println("Rows: " + rows);
assertEquals(3, rows.size());
}
use of com.google.cloud.bigquery.storage.v1beta2.ReadSession in project java-bigquery by googleapis.
the class ConnectionImpl method highThroughPutRead.
@VisibleForTesting
BigQueryResult highThroughPutRead(TableId destinationTable, long totalRows, Schema schema, BigQueryResultStats stats) {
try {
if (bqReadClient == null) {
// if the read client isn't already initialized. Not thread safe.
bqReadClient = BigQueryReadClient.create();
}
String parent = String.format("projects/%s", destinationTable.getProject());
String srcTable = String.format("projects/%s/datasets/%s/tables/%s", destinationTable.getProject(), destinationTable.getDataset(), destinationTable.getTable());
// Read all the columns if the source table (temp table) and stream the data back in Arrow
// format
ReadSession.Builder sessionBuilder = ReadSession.newBuilder().setTable(srcTable).setDataFormat(DataFormat.ARROW);
CreateReadSessionRequest.Builder builder = CreateReadSessionRequest.newBuilder().setParent(parent).setReadSession(sessionBuilder).setMaxStreamCount(// Currently just one stream is allowed
1);
ReadSession readSession = bqReadClient.createReadSession(builder.build());
bufferRow = new LinkedBlockingDeque<>(getBufferSize());
Map<String, Integer> arrowNameToIndex = new HashMap<>();
// deserialize and populate the buffer async, so that the client isn't blocked
processArrowStreamAsync(readSession, bufferRow, new ArrowRowReader(readSession.getArrowSchema(), arrowNameToIndex), schema);
logger.log(Level.INFO, "\n Using BigQuery Read API");
return new BigQueryResultImpl<BigQueryResultImpl.Row>(schema, totalRows, bufferRow, stats);
} catch (IOException e) {
throw BigQueryException.translateAndThrow(e);
}
}
use of com.google.cloud.bigquery.storage.v1beta2.ReadSession in project beam by apache.
the class BigQueryIOStorageQueryTest method doReadFromBigQueryIO.
private void doReadFromBigQueryIO(boolean templateCompatibility) throws Exception {
TableReference sourceTableRef = BigQueryHelpers.parseTableSpec("project:dataset.table");
fakeDatasetService.createDataset(sourceTableRef.getProjectId(), sourceTableRef.getDatasetId(), "asia-northeast1", "Fake plastic tree^H^H^H^Htables", null);
fakeDatasetService.createTable(new Table().setTableReference(sourceTableRef).setLocation("asia-northeast1"));
Table queryResultTable = new Table().setSchema(TABLE_SCHEMA).setNumBytes(0L);
String encodedQuery = FakeBigQueryServices.encodeQueryResult(queryResultTable);
fakeJobService.expectDryRunQuery(options.getProject(), encodedQuery, new JobStatistics().setQuery(new JobStatistics2().setTotalBytesProcessed(1024L * 1024L).setReferencedTables(ImmutableList.of(sourceTableRef))));
ReadSession readSession = ReadSession.newBuilder().setName("readSessionName").setAvroSchema(AvroSchema.newBuilder().setSchema(AVRO_SCHEMA_STRING)).addStreams(ReadStream.newBuilder().setName("streamName")).setDataFormat(DataFormat.AVRO).build();
ReadRowsRequest expectedReadRowsRequest = ReadRowsRequest.newBuilder().setReadStream("streamName").build();
List<GenericRecord> records = Lists.newArrayList(createRecord("A", 1, AVRO_SCHEMA), createRecord("B", 2, AVRO_SCHEMA), createRecord("C", 3, AVRO_SCHEMA), createRecord("D", 4, AVRO_SCHEMA));
List<ReadRowsResponse> readRowsResponses = Lists.newArrayList(createResponse(AVRO_SCHEMA, records.subList(0, 2), 0.0, 0.500), createResponse(AVRO_SCHEMA, records.subList(2, 4), 0.5, 0.875));
//
// Note that since the temporary table name is generated by the pipeline, we can't match the
// expected create read session request exactly. For now, match against any appropriately typed
// proto object.
//
StorageClient fakeStorageClient = mock(StorageClient.class, withSettings().serializable());
when(fakeStorageClient.createReadSession(any())).thenReturn(readSession);
when(fakeStorageClient.readRows(expectedReadRowsRequest, "")).thenReturn(new FakeBigQueryServerStream<>(readRowsResponses));
BigQueryIO.TypedRead<KV<String, Long>> typedRead = BigQueryIO.read(new ParseKeyValue()).fromQuery(encodedQuery).withMethod(Method.DIRECT_READ).withTestServices(new FakeBigQueryServices().withDatasetService(fakeDatasetService).withJobService(fakeJobService).withStorageClient(fakeStorageClient));
if (templateCompatibility) {
typedRead = typedRead.withTemplateCompatibility();
}
PCollection<KV<String, Long>> output = p.apply(typedRead);
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.ReadSession 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.ReadSession in project beam by apache.
the class BigQueryIOStorageReadTest method testReadFromStreamSourceArrow.
@Test
public void testReadFromStreamSourceArrow() throws Exception {
ReadSession readSession = ReadSession.newBuilder().setName("readSession").setArrowSchema(ArrowSchema.newBuilder().setSerializedSchema(serializeArrowSchema(ARROW_SCHEMA)).build()).setDataFormat(DataFormat.ARROW).build();
ReadRowsRequest expectedRequest = ReadRowsRequest.newBuilder().setReadStream("readStream").build();
List<String> names = Arrays.asList("A", "B", "C");
List<Long> values = Arrays.asList(1L, 2L, 3L);
List<ReadRowsResponse> responses = Lists.newArrayList(createResponseArrow(ARROW_SCHEMA, names.subList(0, 2), values.subList(0, 2), 0.0, 0.50), createResponseArrow(ARROW_SCHEMA, names.subList(2, 3), values.subList(2, 3), 0.5, 0.75));
StorageClient fakeStorageClient = mock(StorageClient.class);
when(fakeStorageClient.readRows(expectedRequest, "")).thenReturn(new FakeBigQueryServerStream<>(responses));
BigQueryStorageStreamSource<TableRow> streamSource = BigQueryStorageStreamSource.create(readSession, ReadStream.newBuilder().setName("readStream").build(), TABLE_SCHEMA, new TableRowParser(), TableRowJsonCoder.of(), new FakeBigQueryServices().withStorageClient(fakeStorageClient));
List<TableRow> rows = new ArrayList<>();
BoundedReader<TableRow> reader = streamSource.createReader(options);
for (boolean hasNext = reader.start(); hasNext; hasNext = reader.advance()) {
rows.add(reader.getCurrent());
}
System.out.println("Rows: " + rows);
assertEquals(3, rows.size());
}
Aggregations