While exporting SELECT statement result to BigQuery only empty table is created

Question

I am trying to export a select statement results to another table as a permanent storage. But, whenenver that new table is created it is schemaless. When I try to query that result table an error is shown:

Table project-id.dataset_name.temp_table does not have a schema.

Here is my code to export result from SELECT statement to a temporary tab

def query_to_table():
    service_account_info = {}  # account info

    credentials = Credentials.from_service_account_info(
        service_account_info)
    client = bigquery.Client(
        project=service_account_info.get("project_id"),
        credentials=credentials)

    query = """
            SELECT
                a,
                b

            FROM `project.dataset.table`

            WHERE a NOT IN ('error', 'warning')
        """

    destination_dataset = client.dataset("abc_123") #this is another dataset
    destination_table = destination_dataset.table("temp_table") # destination table

    try:
        client.get_table(destination_table)
        client.delete_table(destination_table)
    except Exception as e:
        # Some logging
        pass

    client.create_table(Table(destination_table))

    # Execute the job and save to table
    job_config = bigquery.QueryJobConfig()
    job_config.allow_large_results = True
    job_config.use_legacy_sql = False
    job_config.destination = destination_table
    job_config.dry_run = True

    query_job = client.query(query, job_config=job_config)

    # Wait till the job done
    while not query_job.done():
        time.sleep(1)

    logging.info(f"Processed {query_job.total_bytes_processed} bytes.")

    return destination_table

Where is the mistake? Is there any API changes from the Google Cloud side? Because this script was working one month earlier.

Please help.

Answer 1

Damn! I just figured it out, it was because I set the dry_run to True.

According to this: https://stackoverflow.com/a/28355802/4494547, if dry_run is set to True, it just evaluates the Query without actually running the job.

Took me 5 hours busting my head. :(