Pub/Sub Avro to BigQuery template

The Pub/Sub Avro to BigQuery template is a streaming pipeline that ingests Avro data from a Pub/Sub subscription into a BigQuery table. Any errors which occur while writing to the BigQuery table are streamed into a Pub/Sub unprocessed topic.

Pipeline requirements

• The input Pub/Sub subscription must exist.
• The schema file for the Avro records must exist on Cloud Storage.
• The unprocessed Pub/Sub topic must exist.
• The output BigQuery dataset must exist.

Template parameters

Parameter	Description
schemaPath	The Cloud Storage location of the Avro schema file. For example, gs://path/to/my/schema.avsc.
inputSubscription	The Pub/Sub input subscription to read from. For example, projects/<project>/subscriptions/<subscription>.
outputTopic	The Pub/Sub topic to use for unprocessed records. For example, projects/<project-id>/topics/<topic-name>.
outputTableSpec	The BigQuery output table location. For example, <my-project>:<my-dataset>.<my-table>. Depending on the createDisposition specified, the output table may be created automatically using the user provided Avro schema.
writeDisposition	(Optional) The BigQuery WriteDisposition. For example, WRITE_APPEND, WRITE_EMPTY or WRITE_TRUNCATE. Default: WRITE_APPEND
createDisposition	(Optional) The BigQuery CreateDisposition. For example, CREATE_IF_NEEDED, CREATE_NEVER. Default: CREATE_IF_NEEDED

Run the template

Console

1. Go to the Dataflow Create job from template page.
Go to Create job from template
2. In the Job name field, enter a unique job name.
3. Optional: For Regional endpoint, select a value from the drop-down menu. The default regional endpoint is us-central1.

   For a list of regions where you can run a Dataflow job, see Dataflow locations.

4. From the Dataflow template drop-down menu, select the Pub/Sub Avro to BigQuery template.
5. In the provided parameter fields, enter your parameter values.
6. Click Run job.

gcloud

Note: To use the Google Cloud CLI to run flex templates, you must have Google Cloud CLI version 284.0.0 or later.

In your shell or terminal, run the template:

gcloud dataflow flex-template run JOB_NAME \
    --region=REGION_NAME \
    --template-file-gcs-location=gs://dataflow-templates-REGION_NAME/VERSION/flex/PubSub_Avro_to_BigQuery \
    --parameters \
schemaPath=SCHEMA_PATH,\
inputSubscription=SUBSCRIPTION_NAME,\
outputTableSpec=BIGQUERY_TABLE,\
outputTopic=DEADLETTER_TOPIC
  

Replace the following:

• JOB_NAME: a unique job name of your choice
• REGION_NAME: the regional endpoint where you want to deploy your Dataflow job—for example, us-central1
• VERSION: the version of the template that you want to use

  You can use the following values:

  • latest to use the latest version of the template, which is available in the non-dated parent folder in the bucket— gs://dataflow-templates-REGION_NAME/latest/
  • the version name, like 2023-04-18-00_RC00, to use a specific version of the template, which can be found nested in the respective dated parent folder in the bucket— gs://dataflow-templates-REGION_NAME/
  Caution: The latest version of templates might update with breaking changes. Your production environments should use templates kept in the most recent dated parent folder to prevent these breaking changes from affecting your production workflows.
• SCHEMA_PATH: the Cloud Storage path to the Avro schema file (for example, gs://MyBucket/file.avsc)
• SUBSCRIPTION_NAME: the Pub/Sub input subscription name
• BIGQUERY_TABLE: the BigQuery output table name
• DEADLETTER_TOPIC: the Pub/Sub topic to use for the unprocessed queue

API

To run the template using the REST API, send an HTTP POST request. For more information on the API and its authorization scopes, see projects.templates.launch.

POST https://dataflow.googleapis.com/v1b3/projects/PROJECT_ID/locations/LOCATION/flexTemplates:launch
{
   "launch_parameter": {
      "jobName": "JOB_NAME",
      "containerSpecGcsPath": "gs://dataflow-templates-LOCATION/VERSION/flex/PubSub_Avro_to_BigQuery",
      "parameters": {
          "schemaPath": "SCHEMA_PATH",
          "inputSubscription": "SUBSCRIPTION_NAME",
          "outputTableSpec": "BIGQUERY_TABLE",
          "outputTopic": "DEADLETTER_TOPIC"
      }
   }
}
  

Replace the following:

• JOB_NAME: a unique job name of your choice
• LOCATION: the regional endpoint where you want to deploy your Dataflow job—for example, us-central1
• VERSION: the version of the template that you want to use

  You can use the following values:

  • latest to use the latest version of the template, which is available in the non-dated parent folder in the bucket— gs://dataflow-templates-REGION_NAME/latest/
  • the version name, like 2023-04-18-00_RC00, to use a specific version of the template, which can be found nested in the respective dated parent folder in the bucket— gs://dataflow-templates-REGION_NAME/
  Caution: The latest version of templates might update with breaking changes. Your production environments should use templates kept in the most recent dated parent folder to prevent these breaking changes from affecting your production workflows.
• SCHEMA_PATH: the Cloud Storage path to the Avro schema file (for example, gs://MyBucket/file.avsc)
• SUBSCRIPTION_NAME: the Pub/Sub input subscription name
• BIGQUERY_TABLE: the BigQuery output table name
• DEADLETTER_TOPIC: the Pub/Sub topic to use for the unprocessed queue

Template source code

Java

View on GitHub Feedback

/*
 * Copyright (C) 2020 Google LLC
 *
 * Licensed under the Apache License, Version 2.0 (the "License"); you may not
 * use this file except in compliance with the License. You may obtain a copy of
 * the License at
 *
 *   http://www.apache.org/licenses/LICENSE-2.0
 *
 * Unless required by applicable law or agreed to in writing, software
 * distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
 * WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the
 * License for the specific language governing permissions and limitations under
 * the License.
 */
package com.google.cloud.teleport.v2.templates;

import com.google.cloud.teleport.metadata.Template;
import com.google.cloud.teleport.metadata.TemplateCategory;
import com.google.cloud.teleport.metadata.TemplateParameter;
import com.google.cloud.teleport.v2.common.UncaughtExceptionLogger;
import com.google.cloud.teleport.v2.options.BigQueryCommonOptions.WriteOptions;
import com.google.cloud.teleport.v2.options.BigQueryStorageApiStreamingOptions;
import com.google.cloud.teleport.v2.options.PubsubCommonOptions.ReadSubscriptionOptions;
import com.google.cloud.teleport.v2.options.PubsubCommonOptions.WriteTopicOptions;
import com.google.cloud.teleport.v2.templates.PubsubAvroToBigQuery.PubsubAvroToBigQueryOptions;
import com.google.cloud.teleport.v2.transforms.BigQueryConverters;
import com.google.cloud.teleport.v2.transforms.ErrorConverters;
import com.google.cloud.teleport.v2.utils.BigQueryIOUtils;
import com.google.cloud.teleport.v2.utils.SchemaUtils;
import org.apache.avro.Schema;
import org.apache.avro.generic.GenericRecord;
import org.apache.beam.sdk.Pipeline;
import org.apache.beam.sdk.PipelineResult;
import org.apache.beam.sdk.coders.AvroCoder;
import org.apache.beam.sdk.io.gcp.bigquery.WriteResult;
import org.apache.beam.sdk.io.gcp.pubsub.PubsubIO;
import org.apache.beam.sdk.options.PipelineOptionsFactory;
import org.apache.beam.sdk.options.Validation.Required;
import org.apache.beam.sdk.schemas.transforms.Convert;
import org.apache.beam.sdk.values.Row;

/**
 * A Dataflow pipeline to stream <a href="https://avro.apache.org/">Apache Avro</a> records from
 * Pub/Sub into a BigQuery table.
 *
 * <p>Any persistent failures while writing to BigQuery will be written to a Pub/Sub dead-letter
 * topic.
 *
 * <p>Check out <a
 * href="https://github.com/GoogleCloudPlatform/DataflowTemplates/blob/main/v2/pubsub-binary-to-bigquery/README_PubSub_Avro_to_BigQuery.md">README</a>
 * for instructions on how to use or modify this template.
 */
@Template(
    name = "PubSub_Avro_to_BigQuery",
    category = TemplateCategory.STREAMING,
    displayName = "Pub/Sub Avro to BigQuery",
    description =
        "A streaming pipeline which inserts Avro records from a Pub/Sub subscription into a"
            + " BigQuery table.",
    optionsClass = PubsubAvroToBigQueryOptions.class,
    flexContainerName = "pubsub-avro-to-bigquery",
    documentation =
        "https://cloud.google.com/dataflow/docs/guides/templates/provided/pubsub-avro-to-bigquery",
    contactInformation = "https://cloud.google.com/support")
public final class PubsubAvroToBigQuery {
  /**
   * Validates input flags and executes the Dataflow pipeline.
   *
   * @param args command line arguments to the pipeline
   */
  public static void main(String[] args) {
    UncaughtExceptionLogger.register();

    PubsubAvroToBigQueryOptions options =
        PipelineOptionsFactory.fromArgs(args)
            .withValidation()
            .as(PubsubAvroToBigQueryOptions.class);

    run(options);
  }

  /**
   * Provides custom {@link org.apache.beam.sdk.options.PipelineOptions} required to execute the
   * {@link PubsubAvroToBigQuery} pipeline.
   */
  public interface PubsubAvroToBigQueryOptions
      extends ReadSubscriptionOptions,
          WriteOptions,
          WriteTopicOptions,
          BigQueryStorageApiStreamingOptions {

    @TemplateParameter.GcsReadFile(
        order = 1,
        description = "Cloud Storage path to the Avro schema file",
        helpText = "Cloud Storage path to Avro schema file. For example, gs://MyBucket/file.avsc.")
    @Required
    String getSchemaPath();

    void setSchemaPath(String schemaPath);
  }

  /**
   * Runs the pipeline with the supplied options.
   *
   * @param options execution parameters to the pipeline
   * @return result of the pipeline execution as a {@link PipelineResult}
   */
  private static PipelineResult run(PubsubAvroToBigQueryOptions options) {
    BigQueryIOUtils.validateBQStorageApiOptionsStreaming(options);

    // Create the pipeline.
    Pipeline pipeline = Pipeline.create(options);

    Schema schema = SchemaUtils.getAvroSchema(options.getSchemaPath());

    WriteResult writeResults =
        pipeline
            .apply(
                "Read Avro records",
                PubsubIO.readAvroGenericRecords(schema)
                    .fromSubscription(options.getInputSubscription())
                    .withDeadLetterTopic(options.getOutputTopic()))
            // Workaround for BEAM-12256. Eagerly convert to rows to avoid
            // the RowToGenericRecord function that doesn't handle all data
            // types.
            // TODO: Remove this workaround when a fix for BEAM-12256 is
            // released.
            .apply(Convert.toRows())
            .apply(
                "Write to BigQuery",
                BigQueryConverters.<Row>createWriteTransform(options).useBeamSchema());

    BigQueryIOUtils.writeResultToBigQueryInsertErrors(writeResults, options)
        .apply(
            "Create error payload",
            ErrorConverters.BigQueryInsertErrorToPubsubMessage.<GenericRecord>newBuilder()
                .setPayloadCoder(AvroCoder.of(schema))
                .setTranslateFunction(BigQueryConverters.TableRowToGenericRecordFn.of(schema))
                .build())
        .apply("Write failed records", PubsubIO.writeMessages().to(options.getOutputTopic()));

    // Execute the pipeline and return the result.
    return pipeline.run();
  }
}