[Python BQ] Allow setting a fixed number of Storage API streams (apac…

…he#28592)

* expose num streams option; fix some streaming tests

* clarify docs; address nit

.test-infra/jenkins/job_PostCommit_Python_CrossLanguage_Gcp_Dataflow.groovy

@@ -28,7 +28,7 @@ import static PythonTestProperties.CROSS_LANGUAGE_VALIDATES_RUNNER_PYTHON_VERSIO
 // Collects tests with the @pytest.mark.uses_gcp_java_expansion_service decorator
 PostcommitJobBuilder.postCommitJob('beam_PostCommit_Python_Xlang_Gcp_Dataflow',
     'Run Python_Xlang_Gcp_Dataflow PostCommit', 'Python_Xlang_Gcp_Dataflow (\"Run Python_Xlang_Gcp_Dataflow PostCommit\")', this) {
-      description('Runs end-to-end cross language GCP IO tests on the Dataflow runner.')
+      description('Runs end-to-end cross language GCP IO tests on the Dataflow runner. \"Run Python_Xlang_Gcp_Dataflow PostCommit\"')
 
 
       // Set common parameters.

.test-infra/jenkins/job_PostCommit_Python_CrossLanguage_Gcp_Direct.groovy

@@ -28,7 +28,7 @@ import static PythonTestProperties.CROSS_LANGUAGE_VALIDATES_RUNNER_PYTHON_VERSIO
 // Collects tests with the @pytest.mark.uses_gcp_java_expansion_service decorator
 PostcommitJobBuilder.postCommitJob('beam_PostCommit_Python_Xlang_Gcp_Direct',
     'Run Python_Xlang_Gcp_Direct PostCommit', 'Python_Xlang_Gcp_Direct (\"Run Python_Xlang_Gcp_Direct PostCommit\")', this) {
-      description('Runs end-to-end cross language GCP IO tests on the Direct runner.')
+      description('Runs end-to-end cross language GCP IO tests on the Direct runner. \"Run Python_Xlang_Gcp_Direct PostCommit\"')
 
       // Set common parameters.
       commonJobProperties.setTopLevelMainJobProperties(delegate)

sdks/java/io/google-cloud-platform/src/main/java/org/apache/beam/sdk/io/gcp/bigquery/providers/BigQueryStorageWriteApiSchemaTransformProvider.java

@@ -176,6 +176,13 @@ public void validate() {
             !Strings.isNullOrEmpty(this.getErrorHandling().getOutput()),
             invalidConfigMessage + "Output must not be empty if error handling specified.");
       }
+
+      if (this.getAutoSharding() != null && this.getAutoSharding()) {
+        checkArgument(
+            this.getNumStreams() == 0,
+            invalidConfigMessage
+                + "Cannot set a fixed number of streams when auto-sharding is enabled. Please pick only one of the two options.");
+      }
     }
 
     /**
@@ -218,11 +225,17 @@ public static Builder builder() {
     public abstract Boolean getUseAtLeastOnceSemantics();
 
     @SchemaFieldDescription(
-        "This option enables using a dynamically determined number of shards to write to "
+        "This option enables using a dynamically determined number of Storage Write API streams to write to "
             + "BigQuery. Only applicable to unbounded data.")
     @Nullable
     public abstract Boolean getAutoSharding();
 
+    @SchemaFieldDescription(
+        "If set, the Storage API sink will default to using this number of write streams. " +
+            "Only applicable to unbounded data.")
+    @Nullable
+    public abstract Integer getNumStreams();
+
     @SchemaFieldDescription("This option specifies whether and where to output unwritable rows.")
     @Nullable
     public abstract ErrorHandling getErrorHandling();
@@ -243,6 +256,8 @@ public abstract static class Builder {
 
       public abstract Builder setAutoSharding(Boolean autoSharding);
 
+      public abstract Builder setNumStreams(Integer numStreams);
+
       public abstract Builder setErrorHandling(ErrorHandling errorHandling);
 
       /** Builds a {@link BigQueryStorageWriteApiSchemaTransformConfiguration} instance. */
@@ -321,13 +336,19 @@ public PCollectionRowTuple expand(PCollectionRowTuple input) {
       if (inputRows.isBounded() == IsBounded.UNBOUNDED) {
         Long triggeringFrequency = configuration.getTriggeringFrequencySeconds();
         Boolean autoSharding = configuration.getAutoSharding();
-        write =
-            write.withTriggeringFrequency(
-                (triggeringFrequency == null || triggeringFrequency <= 0)
-                    ? DEFAULT_TRIGGERING_FREQUENCY
-                    : Duration.standardSeconds(triggeringFrequency));
-        // use default value true for autoSharding if not configured for STORAGE_WRITE_API
-        if (autoSharding == null || autoSharding) {
+        Integer numStreams = configuration.getNumStreams();
+        // Triggering frequency is only applicable for exactly-once
+        if (!configuration.getUseAtLeastOnceSemantics()) {
+          write =
+              write.withTriggeringFrequency(
+                  (triggeringFrequency == null || triggeringFrequency <= 0)
+                      ? DEFAULT_TRIGGERING_FREQUENCY
+                      : Duration.standardSeconds(triggeringFrequency));
+        }
+        // set num streams if specified, otherwise default to autoSharding
+        if (numStreams > 0) {
+          write = write.withNumStorageWriteApiStreams(numStreams);
+        } else if (autoSharding == null || autoSharding) {
           write = write.withAutoSharding();
         }
       }

sdks/python/apache_beam/io/external/xlang_bigqueryio_it_test.py

@@ -30,12 +30,13 @@
 from hamcrest.core import assert_that as hamcrest_assert
 
 import apache_beam as beam
-from apache_beam.io.external.generate_sequence import GenerateSequence
 from apache_beam.io.gcp.bigquery import StorageWriteToBigQuery
 from apache_beam.io.gcp.bigquery_tools import BigQueryWrapper
 from apache_beam.io.gcp.tests.bigquery_matcher import BigqueryFullResultMatcher
 from apache_beam.io.gcp.tests.bigquery_matcher import BigqueryFullResultStreamingMatcher
+from apache_beam.options.pipeline_options import PipelineOptions
 from apache_beam.testing.test_pipeline import TestPipeline
+from apache_beam.transforms.periodicsequence import PeriodicImpulse
 from apache_beam.utils.timestamp import Timestamp
 
 # Protect against environments where bigquery library is not available.
@@ -99,11 +100,13 @@ class BigQueryXlangStorageWriteIT(unittest.TestCase):
   ALL_TYPES_SCHEMA = (
       "int:INTEGER,float:FLOAT,numeric:NUMERIC,str:STRING,"
       "bool:BOOLEAN,bytes:BYTES,timestamp:TIMESTAMP")
+  _RUNNER = ""
 
   def setUp(self):
     self.test_pipeline = TestPipeline(is_integration_test=True)
     self.args = self.test_pipeline.get_full_options_as_args()
     self.project = self.test_pipeline.get_option('project')
+    _RUNNER = PipelineOptions(self.args).get_all_options()['runner']
 
     self.bigquery_client = BigQueryWrapper()
     self.dataset_id = '%s_%s_%s' % (
@@ -244,8 +247,7 @@ def test_write_with_beam_rows(self):
               table=table_id, expansion_service=self.expansion_service))
     hamcrest_assert(p, bq_matcher)
 
-  def run_streaming(
-      self, table_name, auto_sharding=False, use_at_least_once=False):
+  def run_streaming(self, table_name, num_streams=0, use_at_least_once=False):
     elements = self.ELEMENTS.copy()
     schema = self.ALL_TYPES_SCHEMA
     table_id = '{}:{}.{}'.format(self.project, self.dataset_id, table_name)
@@ -260,33 +262,43 @@ def run_streaming(
         streaming=True,
         allow_unsafe_triggers=True)
 
+    auto_sharding = (num_streams == 0)
     with beam.Pipeline(argv=args) as p:
       _ = (
           p
-          | GenerateSequence(
-              start=0, stop=4, expansion_service=self.expansion_service)
-          | beam.Map(lambda x: elements[x])
+          | PeriodicImpulse(0, 4, 1)
+          | beam.Map(lambda t: elements[t])
           | beam.io.WriteToBigQuery(
               table=table_id,
               method=beam.io.WriteToBigQuery.Method.STORAGE_WRITE_API,
               schema=schema,
+              triggering_frequency=1,
               with_auto_sharding=auto_sharding,
+              num_storage_api_streams=num_streams,
               use_at_least_once=use_at_least_once,
               expansion_service=self.expansion_service))
     hamcrest_assert(p, bq_matcher)
 
-  def test_streaming(self):
-    table = 'streaming'
+  @unittest.skipUnless(
+      "dataflowrunner" in _RUNNER.lower(),
+      "The exactly-once route has the requirement "
+      "`beam:requirement:pardo:on_window_expiration:v1`, "
+      "which is currently only supported by the Dataflow runner.")
+  def test_streaming_with_fixed_num_streams(self):
+    table = 'streaming_fixed_num_streams'
+    self.run_streaming(table_name=table, num_streams=4)
+
+  @unittest.skip(
+      "Streaming to the Storage Write API sink with autosharding is broken "
+      "with Dataflow Runner V2.")
+  def test_streaming_with_auto_sharding(self):
+    table = 'streaming_with_auto_sharding'
     self.run_streaming(table_name=table)
 
   def test_streaming_with_at_least_once(self):
-    table = 'streaming'
+    table = 'streaming_with_at_least_once'
     self.run_streaming(table_name=table, use_at_least_once=True)
 
-  def test_streaming_with_auto_sharding(self):
-    table = 'streaming_with_auto_sharding'
-    self.run_streaming(table_name=table, auto_sharding=True)
-
 
 if __name__ == '__main__':
   logging.getLogger().setLevel(logging.INFO)

sdks/python/apache_beam/io/gcp/bigquery.py

@@ -1869,6 +1869,7 @@ def __init__(
       # TODO(https://github.com/apache/beam/issues/20712): Switch the default
       # when the feature is mature.
       with_auto_sharding=False,
+      num_storage_api_streams=0,
       ignore_unknown_columns=False,
       load_job_project_id=None,
       max_insert_payload_size=MAX_INSERT_PAYLOAD_SIZE,
@@ -2018,6 +2019,9 @@ def __init__(
         determined number of shards to write to BigQuery. This can be used for
         all of FILE_LOADS, STREAMING_INSERTS, and STORAGE_WRITE_API. Only
         applicable to unbounded input.
+      num_storage_api_streams: Specifies the number of write streams that the
+        Storage API sink will use. This parameter is only applicable when
+        writing unbounded data.
       ignore_unknown_columns: Accept rows that contain values that do not match
         the schema. The unknown values are ignored. Default is False,
         which treats unknown values as errors. This option is only valid for
@@ -2060,6 +2064,7 @@ def __init__(
     self.use_at_least_once = use_at_least_once
     self.expansion_service = expansion_service
     self.with_auto_sharding = with_auto_sharding
+    self._num_storage_api_streams = num_storage_api_streams
     self.insert_retry_strategy = insert_retry_strategy
     self._validate = validate
     self._temp_file_format = temp_file_format or bigquery_tools.FileFormat.JSON
@@ -2259,6 +2264,7 @@ def find_in_nested_dict(schema):
               triggering_frequency=triggering_frequency,
               use_at_least_once=self.use_at_least_once,
               with_auto_sharding=self.with_auto_sharding,
+              num_storage_api_streams=self._num_storage_api_streams,
               expansion_service=self.expansion_service))
 
       if is_rows:
@@ -2521,6 +2527,7 @@ def __init__(
       triggering_frequency=0,
       use_at_least_once=False,
       with_auto_sharding=False,
+      num_storage_api_streams=0,
       expansion_service=None):
     """Initialize a StorageWriteToBigQuery transform.
 
@@ -2558,6 +2565,7 @@ def __init__(
     self._triggering_frequency = triggering_frequency
     self._use_at_least_once = use_at_least_once
     self._with_auto_sharding = with_auto_sharding
+    self._num_storage_api_streams = num_storage_api_streams
     self._expansion_service = (
         expansion_service or _default_io_expansion_service())
     self.schematransform_config = SchemaAwareExternalTransform.discover_config(
@@ -2569,6 +2577,7 @@ def expand(self, input):
         expansion_service=self._expansion_service,
         rearrange_based_on_discovery=True,
         autoSharding=self._with_auto_sharding,
+        numStreams=self._num_storage_api_streams,
         createDisposition=self._create_disposition,
         table=self._table,
         triggeringFrequencySeconds=self._triggering_frequency,

website/www/site/content/en/documentation/io/built-in/google-bigquery.md

@@ -788,7 +788,7 @@ BigQuery Storage Write API for Python SDK currently has some limitations on supp
 {{< paragraph class="language-py" >}}
 **Note:** If you want to run WriteToBigQuery with Storage Write API from the source code, you need to run `./gradlew :sdks:java:io:google-cloud-platform:expansion-service:build` to build the expansion-service jar. If you are running from a released Beam SDK, the jar will already be included.
 
-**Note:** Auto sharding is not currently supported for Python's Storage Write API.
+**Note:** Auto sharding is not currently supported for Python's Storage Write API exactly-once mode on DataflowRunner.
 
 {{< /paragraph >}}