Using results of completed jobs

Sometimes you need to process the job outputs or use them to run another job. To do this, download them using the DataSphere Jobs SDK download_job_files function or feed them into another job using the DAG Airflow operator.

Downloading via DataSphere CLIDownloading via DataSphere CLI

If the total size of job results exceeds the maximum total size of files you can download via DataSphere CLI, only the files with the total size not exceeding this limit will be downloaded.

To download job results into the working directory using DataSphere CLI, run this command by substituting the job ID:

datasphere project job download-files --id <job_ID>

The function has the following optional parameters:

• with_logs: Downloads logs. The default value is false.
• with_diagnostics: Downloads diagnostic files. The default value is false.
• output_dir: File download directory. By default, this is your working directory.

To use the downloaded data in another job, specify the data in the input parameter in the config.yaml file.

Downloading via DAG AirflowDownloading via DAG Airflow

You can use a DAG operator to download the output file of one job (result.txt) and provide it as the input file (INPUT_DATA) for another:

from typing import Dict

from airflow.decorators import dag, task
import pendulum

from datasphere import SDK

now = pendulum.now()

@dag(dag_id='output_files_for_other_job', start_date=now, schedule="@daily", catchup=False)
def run():
    @task(task_id='fork_job')
    def fork_job(files_job_id: str, fork_source_job_id: str):
        sdk = SDK()
        sdk.download_job_files(files_job_id)
        job = sdk.fork_job(fork_source_job_id, vars={'INPUT_DATA': 'result.txt'})
        job.wait()

    fork_job('<ID_of_job_to_download>', '<ID_of_job_to_run>')

Useful linksUseful links

• DataSphere CLI