RDA (DigitalGlobe Raster Data Access)

Driver short name


New in version 2.3.

Build dependencies


This driver can connect to DigitalGlobe RDA REST API. GDAL/OGR must be built with Curl support in order for the RDA driver to be compiled.

The driver retrieves metadata on graphs and fetches the raster by tiles. Data types byte, uint16, int16, uint32, int32, float32 and float64 are supported.

Any valid graph or template is supported via the DigitalGlobe RDA REST API.

There is no support for overviews.

Driver capabilities

Supports Georeferencing

This driver supports georeferencing

Dataset name syntax

The minimal syntax to open a datasource is :

{"graphId":"some_value", "nodeId": "another_value"}


{"templateId":"some_value", "parameters": { "someparam": "someparamval"}}

So a JSon serialized document with 2 attributes graph-id and node-id.

Those values can for example be retrieved from graphs built by GraphStudio.

Connection String options (optional)

  • "options": {"delete-on-close": false}

    can be added to the JSon document to request that cached tiles and metadata are not destroyed at dataset closing. The default, if not specified, is true.

  • "options": {"max-connections": 32}

    can be added to the JSon document to request that cached tiles be fetched using a maximum number of concurrent connections. The default, if not specified, is equal to 8 * number of CPUs.

  • "options": {"advise-read": false}

    can be added to the JSon document to request advise read not be used when reading the dataset. The default, if not specified, is true.


Access to the API requires an authentication token. For that, 2 parameters (username, password) must be provided to the driver. They can be retrieved from the below configuration options, or from the ~/.gbdx-config file.

The access token will be cached in ~/.gdal/rda_cache/authentication.json and reused from there until its expiration period is reached.

Configuration options

The following configuration options are available :

  • GBDX_AUTH_URL=value: To specify the OAuth authentication endpoint. Defaults to https://geobigdata.io/auth/v1/oauth/token/. If not specified, the auth_url parameter from ~/.gbdx-config will be used if it exists.

  • GBDX_RDA_API_URL=value: To specify the RDA API endpoint. Defaults to https://rda.geobigdata.io/v1. If not specified, the rda_api_url parameter from ~/.gbdx-config will be used if it exists.

  • GBDX_USERNAME=value: To specify the OAuth user name needed to get to an authentication token. If not specified, the user_name parameter from ~/.gbdx-config must be set.

  • GBDX_PASSWORD=value: To specify the OAuth user name needed to get to an authentication token. If not specified, the password parameter from ~/.gbdx-config must be set.

~/.gbdx-config file

This file may be created in the home directory of the user (value of the $HOME environment variable on Unix, $USERPROFILE on Windows). It can contain values from the above configuration options.

auth_url = https://geobigdata.io/auth/v1/oauth/token/ (optional)
rda_api_url = https://rda.geobigdata.io/v1 (optional)
user_name = value (required)
user_password = value (required)


By default, the authentication token is cached in the ~/.gdal/rda_cache directory. This directory may be changed with the RDA_CACHE_DIR configuration option. By default, dataset metadata and tiles are temporarily cached in ~/.gdal/rda_cache/{graph-id}/{node-id}, and deleted on dataset closing, unless

"options": {"delete-on-close": false}

is found in the dataset name.

Open Options

By default, the number of concurrent downloads will be 8*number of CPUs up to a maximum of 64. The maximum number of concurrent connections can be configured by the MAXCONNECT option


  • Display metadata, and keep it cached:

    gdalinfo '{"graphId":"832050eb7d271d8704c8889369ee0a8a1da82acdee1b20e1700b6d053e94d1fe","nodeId":"Orthorectify_hko89y", "options": {"delete-on-close": false}}'
    Driver: RDA/DigitalGlobe Raster Data Access driver
    Files: none associated
    Size is 9911, 7084
    Coordinate System is:
    GEOGCS["WGS 84",
            SPHEROID["WGS 84",6378137,298.257223563,
    Origin = (-84.183163638386631,33.835018117204456)
    Pixel Size = (0.000020885734819,-0.000020885734819)
      GSD=2.325 m
      SENSOR_NAME=8-band (Coastal, Blue, Green, Yellow, Red, Red-edge, NIR1, NIR2) Multispectral
    Image Structure Metadata:
    Corner Coordinates:
    Upper Left  ( -84.1831636,  33.8350181)
    Lower Left  ( -84.1831636,  33.6870636)
    Upper Right ( -83.9761651,  33.8350181)
    Lower Right ( -83.9761651,  33.6870636)
    Center      ( -84.0796644,  33.7610408)
    Band 1 Block=256x256 Type=UInt16, ColorInterp=Undefined
    Band 2 Block=256x256 Type=UInt16, ColorInterp=Blue
    Band 3 Block=256x256 Type=UInt16, ColorInterp=Green
    Band 4 Block=256x256 Type=UInt16, ColorInterp=Yellow
    Band 5 Block=256x256 Type=UInt16, ColorInterp=Red
    Band 6 Block=256x256 Type=UInt16, ColorInterp=Undefined
    Band 7 Block=256x256 Type=UInt16, ColorInterp=Undefined
    Band 8 Block=256x256 Type=UInt16, ColorInterp=Undefined
  • Extract a subwindow from a dataset:

    gdal_translate -srcwin 1000 2000 500 500 '{"graphId":"832050eb7d271d8704c8889369ee0a8a1da82acdee1b20e1700b6d053e94d1fe","nodeId":"Orthorectify_hko89y"}' out.tif
  • Materialize a dataset specifying a custom number of concurrent connections:

    gdal_translate -oo MAXCONNECT=96 '{"graphId":"832050eb7d271d8704c8889369ee0a8a1da82acdee1b20e1700b6d053e94d1fe","nodeId":"Orthorectify_hko89y"}' out.tif
  • Materialize a dataset from a template:

    gdal_translate '{"templateId": "sample", "parameters": { "imageId": "afa56b05-35ad-47d1-bc7f-3e23d220482d"}}' out.tif