Cloud Storage transfers

The BigQuery Data Transfer Service for Cloud Storage allows you to schedule recurring data loads from Cloud Storage to BigQuery.

Before you begin

Before you create a Cloud Storage transfer:

Limitations

Recurring transfers from Cloud Storage to BigQuery are subject to the following limitations:

  • All files matching the patterns defined by either a wildcard or by runtime parameters for your transfer must share the same schema you defined for the destination table, or the transfer will fail. Table schema changes between runs also cause the transfer to fail.
  • Source files in Cloud Storage need to be at least one hour old to get picked up by the transfer.
  • Because Cloud Storage objects can be versioned, it's important to note that archived Cloud Storage objects are not supported for BigQuery transfers. Objects must be live to be transferred.
  • Unlike individual loads of data from Cloud Storage to BigQuery, for ongoing transfers you need to create the destination table and its schema in advance of setting up the transfer. BigQuery cannot create the table as part of the recurring data transfer process.
  • Transfers from Cloud Storage are always triggered with the WRITE_APPEND preference which appends data to the destination table. See configuration.load.writeDisposition under load configuration for additional details.
  • If your dataset's location is set to a value other than US, the regional or multi-regional Cloud Storage bucket must be in the same region as the dataset.

Depending on the format of your Cloud Storage source data, there may be additional limitations. For more information, see:

Required permissions

When you load data into BigQuery, you need project or dataset- level permissions that allow you to load data into new or existing BigQuery tables and partitions. If you are loading data from Cloud Storage, you'll also need access to the bucket that contains your data. Ensure that you have the following required permissions:

  • BigQuery: bigquery.transfers.update permissions are required to create the scheduled transfer. The bigquery.admin predefined, project-level IAM role includes bigquery.transfers.update permissions. For more information on IAM roles in BigQuery, see Access Control.
  • Cloud Storage: storage.objects.get permissions are required at the project level or on the individual bucket. If you are using a URI wildcard, you must also have storage.objects.list permissions. If you would like to delete the source files after each successful transfer, you also need storage.objects.delete permissions. The storage.objectAdmin predefined, project-level IAM role includes all of these permissions.

Setting up a Cloud Storage transfer

To create a Cloud Storage transfer in BigQuery Data Transfer Service:

Console

  1. Go to the BigQuery web UI.

    Go to the BigQuery web UI

  2. Click Transfers.

  3. Click Create.

  4. On the Create Transfer page:

    • In the Source type section, for Source, choose Cloud Storage.

      Transfer source

    • In the Transfer config name section, for Display name, enter a name for the transfer such as My Transfer. The transfer name can be any value that allows you to easily identify the transfer if you need to modify it later.

      Transfer name

    • In the Schedule options section, for Schedule, leave the default value (Start now) or click Start at a set time.

      • For Repeats, choose an option for how often to run the transfer.
        • Daily (default)
        • Weekly
        • Monthly
        • Custom
        • On-demand
      • For Start date and run time, enter the date and time to start the transfer. If you choose Start now, this option is disabled.

        Transfer schedule

    • In the Destination settings section, for Destination dataset, choose the dataset you created to store your data.

      Transfer dataset

    • In the Data source details section:

      • For Destination table, enter the name of your destination table. The destination table must follow the table naming rules. Destination table names also support parameters.
      • For Cloud Storage URI, enter the Cloud Storage URI. Wildcards and parameters are supported.
      • For Delete source files after transfer, check the box if you want to delete the source files after each successful transfer. Delete jobs are best effort. Delete jobs do not retry if the first effort to delete the source files fails.
      • In the Transfer Options section:

        • Under All Formats:
          • For Number of errors allowed, enter the maximum number of bad records that BigQuery can ignore when running the job. If the number of bad records exceeds this value, an 'invalid' error is returned in the job result, and the job fails. The default value is 0.
        • Under JSON, CSV:
          • For Ignore unknown values, check the box if you want the transfer to drop data that does not fit the destination table's schema.
        • Under CSV:

          • For Field delimiter, enter the character that separates fields. The default value is a comma.
          • For Header rows to skip, enter the number of header rows in the source file(s) if you don't want to import them. The default value is 0.
          • For Allow quoted newlines, check the box if you want to allow newlines within quoted fields.
          • For Allow jagged rows, check the box if you want to allow the transfer of rows with missing NULLABLE columns.

      Cloud Storage source details

    • (Optional) In the Notification options section:

      • Click the toggle to enable email notifications. When you enable this option, the transfer administrator receives an email notification when a transfer run fails.
      • For Select a Cloud Pub/Sub topic, choose your topic name or click Create a topic. This option configures Cloud Pub/Sub run notifications for your transfer. Transfer run notifications are currently in alpha.
  5. Click Save.

Classic UI

  1. Go to the BigQuery web UI.

    Go to the BigQuery web UI

  2. Click Transfers.

  3. Click Add Transfer.

  4. On the New Transfer page:

    • For Source, choose Cloud Storage.
    • For Display name, enter a name for the transfer such as My Transfer. The display name can be any value that allows you to easily identify the transfer if you need to modify it later.
    • (Optional) For Schedule, you can leave the default value of Daily (every 24 hours, based on creation time), or click Edit to change the time. You can also change the interval to Weekly, Monthly, or Custom. When selecting Custom, a Cron-like time specification is expected, for example every 12 hours. The shortest allowed period is 12 hours. See the schedule field under TransferConfig for additional valid API values.

      transfer schedule

    • For Destination dataset, choose the appropriate dataset.

    • For Destination table, enter the name of your destination table. The destination table must follow the table naming rules. Destination table names also support parameters.

    • For Cloud Storage URI, enter the Cloud Storage URI. Wildcards and parameters are supported.

    • For Delete source files after transfer, check the box if you want to delete the source files after each successful transfer. Delete jobs are best effort. Delete jobs do not retry if the first effort to delete the source files fails.

    • For File format, select the type of files you wish to transfer.

    • Under the Transfer Options - All Formats section:

      • For Number of errors allowed, enter the maximum number of bad records that BigQuery can ignore when running the job. If the number of bad records exceeds this value, an 'invalid' error is returned in the job result, and the job fails. The default value is 0.
    • Under the Transfer Options - JSON, CSV section:

      • For Ignore unknown values, check the box if you want the transfer to drop data that does not fit the destination table's schema.
    • Under the Transfer Options - CSV section:

      • For Field delimiter, enter the character that separates fields. The default value is a comma.
      • For Header rows to skip, enter the number of header rows in the source file(s) if you don't want to import them. The default value is 0.
      • For Allow quoted newlines, check the box if you want to allow newlines within quoted fields.
      • For Allow jagged rows, check the box if you want to allow the transfer of rows with missing NULLABLE columns.

      New Cloud Storage transfer

    • (Optional) Expand the Advanced section and configure run notifications for your transfer. Transfer run notifications are currently in Alpha.

    • For Cloud Cloud Pub/Sub topic, enter your Cloud Pub/Sub topic name, for example, projects/myproject/topics/mytopic.

    • Check Send email notifications to allow email notifications of transfer run failures.

      Cloud Cloud Pub/Sub topic

  5. Click Add.

Command-line

Enter the bq mk command and supply the transfer creation flag — --transfer_config. The following flags are also required:

  • --data_source
  • --display_name
  • --target_dataset
  • --params

    bq mk --transfer_config --project_id=[PROJECT_ID] \
    --data_source=[DATA_SOURCE] --display_name=[NAME] \
    --target_dataset=[DATASET] --params='[PARAMETERS]'
    

Where:

  • --project_id is your project ID. If --project_id isn't supplied to specify a particular project, the default project is used.
  • --data_source is the data source — google_cloud_storage.
  • --display_name is the display name for the transfer configuration. The transfer name can be any value that allows you to easily identify the transfer if you need to modify it later.
  • --target_dataset is the target dataset for the transfer configuration.
  • --params contains the parameters for the created transfer configuration in JSON format. For example: --params='{"param":"param_value"}'.
    • For Cloud Storage, you must supply the data_path_template, the destination_table_name_template and the file_format parameters. data_path_template is the Cloud Storage URI that contains your files to be transferred, which can include one wildcard. The destination_table_name_template is the name of your destination table. For file_format, indicate the type of files you wish to transfer: CSV, JSON, AVRO, PARQUET, or ORC. The default value is CSV.
    • For all file_format values, you can include the optional param max_bad_records. The default value is 0.
    • For the JSON or CSV values in file_format, you can include the optional param ignore_unknown_values. This param will be ignored if you haven't selected CSV or JSON for the file_format.
    • For CSV file_format, you can include the optional param field_delimiter for the character that separates fields. The default value is a comma. This param will be ignored if you haven't selected CSV for the file_format.
    • For CSV file_format, you can include the optional param skip_leading_rows to indicate header rows you don't want to import. The default value is 0. This param will be ignored if you haven't selected CSV for the file_format.
    • For CSV file_format, you can include the optional param allow_quoted_newlines if you want to allow newlines within quoted fields. This param will be ignored if you haven't selected CSV for the file_format.
    • For CSV file_format, you can include the optional param allow_jagged_rows if you want to accept rows that are missing trailing optional columns. The missing values will be filled in with NULLs. This param will be ignored if you haven't selected CSV for the file_format.
    • Optional param delete_source_files will delete the source files after each successful transfer. (Delete jobs do not retry if the first effort to delete the source files fails.) The default value for the delete_source_files is false.

For example, the following command creates a Cloud Storage transfer named My Transfer using a data_path_template value of gs://mybucket/myfile/*.csv, target dataset mydataset, and file_format CSV. This example includes non-default values for the optional params associated with the CSV file_format.

The transfer is created in the default project:

bq mk --transfer_config --target_dataset=mydataset \
--display_name='My Transfer' \
--params='{"data_path_template":"gs://mybucket/myfile/*.csv",
"destination_table_name_template":"MyTable",
"file_format":"CSV",
"max_bad_records":"1",
"ignore_unknown_values":"true",
"field_delimiter":"|",
"skip_leading_rows":"1",
"allow_quoted_newlines":"true",
"allow_jagged_rows":"false",
"delete_source_files":"true"}' \
--data_source=google_cloud_storage

After running the command, you receive a message like the following:

[URL omitted] Please copy and paste the above URL into your web browser and follow the instructions to retrieve an authentication code.

Follow the instructions and paste the authentication code on the command line.

API

Use the projects.locations.transferConfigs.create method and supply an instance of the TransferConfig resource.

Setting up a refresh run of a transfer

In addition to setting up a recurring transfer from Cloud Storage, you can also set up a refresh run to pick up additional data files.

If the transfer configuration is date-related (parameterized), the Cloud Storage URI is parameterized, or both, you can run the refresh for specific dates.

To set up a refresh transfer:

Console

  1. Open the BigQuery web UI in the GCP Console.

    Go to the BigQuery web UI

  2. Click Transfers.

  3. Click your transfer.

  4. Click the More menu, and then select Refresh Transfer.

  5. In the Schedule a backfill run dialog, choose your Start date and End date. You can use the classic BigQuery web UI to set finer-grained time limits.

    Set refresh dates

Classic UI

  1. Go to the BigQuery web UI.

    Go to the BigQuery web UI

  2. Click Transfers.

  3. Click your transfer.

  4. Click Refresh Transfer.

    Refresh Transfer

  5. In the Start Transfer Runs dialog, choose your Start Time and End Time.

    Set refresh dates

    If your Cloud Storage transfer configuration is not parameterized, you won't be given date options when you click Refresh Transfer. Instead, the refresh will happen immediately.

    Refresh immediately

  6. Click OK.

What's next

Oliko tästä sivusta apua? Kerro mielipiteesi

Palautteen aihe:

Tämä sivu
BigQuery
Tarvitsetko apua? Siirry tukisivullemme.