Use StartJobRun com um AWS SDK ou CLI

PDF

Os exemplos de código a seguir mostram como usar o StartJobRun.

Exemplos de ações são trechos de código de programas maiores e devem ser executados em contexto. É possível ver essa ação em contexto no seguinte exemplo de código:

• Conheça os conceitos básicos

.NET

AWS SDK for .NET

nota

Tem mais sobre GitHub. Encontre o exemplo completo e saiba como configurar e executar no Repositório de exemplos de código da AWS.

    /// <summary>
    /// Start an AWS Glue job run.
    /// </summary>
    /// <param name="jobName">The name of the job.</param>
    /// <returns>A string representing the job run Id.</returns>
    public async Task<string> StartJobRunAsync(
        string jobName,
        string inputDatabase,
        string inputTable,
        string bucketName)
    {
        var request = new StartJobRunRequest
        {
            JobName = jobName,
            Arguments = new Dictionary<string, string>
            {
                {"--input_database", inputDatabase},
                {"--input_table", inputTable},
                {"--output_bucket_url", $"s3://{bucketName}/"}
            }
        };

        var response = await _amazonGlue.StartJobRunAsync(request);
        return response.JobRunId;
    }

• Para obter detalhes da API, consulte StartJobRuna Referência AWS SDK for .NET da API.

C++

SDK para C++

nota

Tem mais sobre GitHub. Encontre o exemplo completo e saiba como configurar e executar no Repositório de exemplos de código da AWS.

        Aws::Client::ClientConfiguration clientConfig;
        // Optional: Set to the AWS Region in which the bucket was created (overrides config file).
        // clientConfig.region = "us-east-1";

    Aws::Glue::GlueClient client(clientConfig);

        Aws::Glue::Model::StartJobRunRequest request;
        request.SetJobName(JOB_NAME);

        Aws::Map<Aws::String, Aws::String> arguments;
        arguments["--input_database"] = CRAWLER_DATABASE_NAME;
        arguments["--input_table"] = tableName;
        arguments["--output_bucket_url"] = Aws::String("s3://") + bucketName + "/";
        request.SetArguments(arguments);

        Aws::Glue::Model::StartJobRunOutcome outcome = client.StartJobRun(request);

        if (outcome.IsSuccess()) {
            std::cout << "Successfully started the job." << std::endl;

            Aws::String jobRunId = outcome.GetResult().GetJobRunId();

            int iterator = 0;
            bool done = false;
            while (!done) {
                ++iterator;
                std::this_thread::sleep_for(std::chrono::seconds(1));
                Aws::Glue::Model::GetJobRunRequest jobRunRequest;
                jobRunRequest.SetJobName(JOB_NAME);
                jobRunRequest.SetRunId(jobRunId);

                Aws::Glue::Model::GetJobRunOutcome jobRunOutcome = client.GetJobRun(
                        jobRunRequest);

                if (jobRunOutcome.IsSuccess()) {
                    const Aws::Glue::Model::JobRun &jobRun = jobRunOutcome.GetResult().GetJobRun();
                    Aws::Glue::Model::JobRunState jobRunState = jobRun.GetJobRunState();

                    if ((jobRunState == Aws::Glue::Model::JobRunState::STOPPED) ||
                        (jobRunState == Aws::Glue::Model::JobRunState::FAILED) ||
                        (jobRunState == Aws::Glue::Model::JobRunState::TIMEOUT)) {
                        std::cerr << "Error running job. "
                                  << jobRun.GetErrorMessage()
                                  << std::endl;
                        deleteAssets(CRAWLER_NAME, CRAWLER_DATABASE_NAME, JOB_NAME,
                                     bucketName,
                                     clientConfig);
                        return false;
                    }
                    else if (jobRunState ==
                             Aws::Glue::Model::JobRunState::SUCCEEDED) {
                        std::cout << "Job run succeeded after  " << iterator <<
                                  " seconds elapsed." << std::endl;
                        done = true;
                    }
                    else if ((iterator % 10) == 0) { // Log status every 10 seconds.
                        std::cout << "Job run status " <<
                                  Aws::Glue::Model::JobRunStateMapper::GetNameForJobRunState(
                                          jobRunState) <<
                                  ". " << iterator <<
                                  " seconds elapsed." << std::endl;
                    }
                }
                else {
                    std::cerr << "Error retrieving job run state. "
                              << jobRunOutcome.GetError().GetMessage()
                              << std::endl;
                    deleteAssets(CRAWLER_NAME, CRAWLER_DATABASE_NAME, JOB_NAME,
                                 bucketName, clientConfig);
                    return false;
                }
            }
        }
        else {
            std::cerr << "Error starting a job. " << outcome.GetError().GetMessage()
                      << std::endl;
            deleteAssets(CRAWLER_NAME, CRAWLER_DATABASE_NAME, JOB_NAME, bucketName,
                         clientConfig);
            return false;
        }

• Para obter detalhes da API, consulte StartJobRuna Referência AWS SDK for C++ da API.

CLI

AWS CLI

Para iniciar a execução de um trabalho

O exemplo de start-job-run a seguir inicia um trabalho.

aws glue start-job-run \
    --job-name my-job

Saída:

{
    "JobRunId": "jr_22208b1f44eb5376a60569d4b21dd20fcb8621e1a366b4e7b2494af764b82ded"
}

Para obter mais informações, consulte Criação de trabalhos no Guia do desenvolvedor do AWS Glue.

• Para obter detalhes da API, consulte StartJobRunem Referência de AWS CLI Comandos.

Java

SDK para Java 2.x

nota

Tem mais sobre GitHub. Encontre o exemplo completo e saiba como configurar e executar no Repositório de exemplos de código da AWS.

    /**
     * Starts a job run in AWS Glue.
     *
     * @param glueClient    the AWS Glue client to use for the job run
     * @param jobName       the name of the Glue job to run
     * @param inputDatabase the name of the input database
     * @param inputTable    the name of the input table
     * @param outBucket     the URL of the output S3 bucket
     * @throws GlueException if there is an error starting the job run
     */
    public static void startJob(GlueClient glueClient, String jobName, String inputDatabase, String inputTable,
                                String outBucket) {
        try {
            Map<String, String> myMap = new HashMap<>();
            myMap.put("--input_database", inputDatabase);
            myMap.put("--input_table", inputTable);
            myMap.put("--output_bucket_url", outBucket);

            StartJobRunRequest runRequest = StartJobRunRequest.builder()
                .workerType(WorkerType.G_1_X)
                .numberOfWorkers(10)
                .arguments(myMap)
                .jobName(jobName)
                .build();

            StartJobRunResponse response = glueClient.startJobRun(runRequest);
            System.out.println("The request Id of the job is " + response.responseMetadata().requestId());

        } catch (GlueException e) {
            throw e;
        }
    }

• Para obter detalhes da API, consulte StartJobRuna Referência AWS SDK for Java 2.x da API.

JavaScript

SDK para JavaScript (v3)

nota

Tem mais sobre GitHub. Encontre o exemplo completo e saiba como configurar e executar no Repositório de exemplos de código da AWS.

const startJobRun = (jobName, dbName, tableName, bucketName) => {
  const client = new GlueClient({});

  const command = new StartJobRunCommand({
    JobName: jobName,
    Arguments: {
      "--input_database": dbName,
      "--input_table": tableName,
      "--output_bucket_url": `s3://${bucketName}/`,
    },
  });

  return client.send(command);
};

• Para obter detalhes da API, consulte StartJobRuna Referência AWS SDK for JavaScript da API.

PHP

SDK para PHP

nota

Tem mais sobre GitHub. Encontre o exemplo completo e saiba como configurar e executar no Repositório de exemplos de código da AWS.

        $jobName = 'test-job-' . $uniqid;

        $databaseName = "doc-example-database-$uniqid";

        $tables = $glueService->getTables($databaseName);

        $outputBucketUrl = "s3://$bucketName";
        $runId = $glueService->startJobRun($jobName, $databaseName, $tables, $outputBucketUrl)['JobRunId'];

    public function startJobRun($jobName, $databaseName, $tables, $outputBucketUrl): Result
    {
        return $this->glueClient->startJobRun([
            'JobName' => $jobName,
            'Arguments' => [
                'input_database' => $databaseName,
                'input_table' => $tables['TableList'][0]['Name'],
                'output_bucket_url' => $outputBucketUrl,
                '--input_database' => $databaseName,
                '--input_table' => $tables['TableList'][0]['Name'],
                '--output_bucket_url' => $outputBucketUrl,
            ],
        ]);
    }

• Para obter detalhes da API, consulte StartJobRuna Referência AWS SDK for PHP da API.

Python

SDK para Python (Boto3)

nota

Tem mais sobre GitHub. Encontre o exemplo completo e saiba como configurar e executar no Repositório de exemplos de código da AWS.

class GlueWrapper:
    """Encapsulates AWS Glue actions."""

    def __init__(self, glue_client):
        """
        :param glue_client: A Boto3 Glue client.
        """
        self.glue_client = glue_client


    def start_job_run(self, name, input_database, input_table, output_bucket_name):
        """
        Starts a job run. A job run extracts data from the source, transforms it,
        and loads it to the output bucket.

        :param name: The name of the job definition.
        :param input_database: The name of the metadata database that contains tables
                               that describe the source data. This is typically created
                               by a crawler.
        :param input_table: The name of the table in the metadata database that
                            describes the source data.
        :param output_bucket_name: The S3 bucket where the output is written.
        :return: The ID of the job run.
        """
        try:
            # The custom Arguments that are passed to this function are used by the
            # Python ETL script to determine the location of input and output data.
            response = self.glue_client.start_job_run(
                JobName=name,
                Arguments={
                    "--input_database": input_database,
                    "--input_table": input_table,
                    "--output_bucket_url": f"s3://{output_bucket_name}/",
                },
            )
        except ClientError as err:
            logger.error(
                "Couldn't start job run %s. Here's why: %s: %s",
                name,
                err.response["Error"]["Code"],
                err.response["Error"]["Message"],
            )
            raise
        else:
            return response["JobRunId"]

• Para obter detalhes da API, consulte a StartJobRunReferência da API AWS SDK for Python (Boto3).

Ruby

SDK para Ruby

nota

Tem mais sobre GitHub. Encontre o exemplo completo e saiba como configurar e executar no Repositório de exemplos de código da AWS.

# The `GlueWrapper` class serves as a wrapper around the AWS Glue API, providing a simplified interface for common operations.
# It encapsulates the functionality of the AWS SDK for Glue and provides methods for interacting with Glue crawlers, databases, tables, jobs, and S3 resources.
# The class initializes with a Glue client and a logger, allowing it to make API calls and log any errors or informational messages.
class GlueWrapper
  def initialize(glue_client, logger)
    @glue_client = glue_client
    @logger = logger
  end

  # Starts a job run for the specified job.
  #
  # @param name [String] The name of the job to start the run for.
  # @param input_database [String] The name of the input database for the job.
  # @param input_table [String] The name of the input table for the job.
  # @param output_bucket_name [String] The name of the output S3 bucket for the job.
  # @return [String] The ID of the started job run.
  def start_job_run(name, input_database, input_table, output_bucket_name)
    response = @glue_client.start_job_run(
      job_name: name,
      arguments: {
        '--input_database': input_database,
        '--input_table': input_table,
        '--output_bucket_url': "s3://#{output_bucket_name}/"
      }
    )
    response.job_run_id
  rescue Aws::Glue::Errors::GlueException => e
    @logger.error("Glue could not start job run #{name}: \n#{e.message}")
    raise
  end

• Para obter detalhes da API, consulte StartJobRuna Referência AWS SDK for Ruby da API.

Rust

SDK para Rust

nota

Tem mais sobre GitHub. Encontre o exemplo completo e saiba como configurar e executar no Repositório de exemplos de código da AWS.

        let job_run_output = glue
            .start_job_run()
            .job_name(self.job())
            .arguments("--input_database", self.database())
            .arguments(
                "--input_table",
                self.tables
                    .first()
                    .ok_or_else(|| GlueMvpError::Unknown("Missing crawler table".into()))?
                    .name(),
            )
            .arguments("--output_bucket_url", self.bucket())
            .send()
            .await
            .map_err(GlueMvpError::from_glue_sdk)?;

        let job = job_run_output
            .job_run_id()
            .ok_or_else(|| GlueMvpError::Unknown("Missing run id from just started job".into()))?
            .to_string();

• Para obter detalhes da API, consulte a StartJobRunreferência da API AWS SDK for Rust.