Handle Open Cloud data store requests

There are a number of considerations around sending requests to and handling responses from the Open Cloud data store APIs.

HTTP headers

Like all Open Cloud endpoints, data store requests that use API key authentication must include the x-api-key header. For more information, see Manage API keys.

  • An incorrect or missing API key results in a 401 error.
  • A valid API key that lacks the correct permissions results in a 403 error.

If a request includes a body, such as a create or update request, include the Content-Type: application/json and Content-Length headers. Most clients include these headers automatically.

Pagination

Data stores often contain many entries, which means your code, particularly when calling List Data Store Entries, must deal with a maximum number of results and pages. To learn more, see Pagination.

Filters

If you use prefixes in the names of your data store entries, you likely want to filter your GET calls, like so:

def list_entries(universe, data_store, filter):

    list_path = f'universes/{universe}/data-stores/{data_store}/entries'

    url = base_url + list_path

    return requests.get(url, params={'filter': f'id.startsWith("global/{filter}")'}, headers={apiHeaderKey: apiKey})

Filter syntax is extremely limited. See the reference documentation for available filters for standard and ordered data stores.

Allow missing flags

The PATCH (update) methods for standard and ordered data stores have an allow_missing parameter. If set to true, this parameter lets you create entries that don't already exist rather than throwing an error. Leave this parameter at its default value of false unless you have a good reason not to; updating an entry that was never created often represents an error in programming logic.

Scopes

Many requests require you to specify a scope, even if it's the default value of global. These scopes are distinct from permission scopes and let you subdivide and organize the entries in a data store.

Rather than creating non-default scopes, however, we strongly recommend using prefixes. To learn more, see Listing and prefixes.

Content-MD5

The Open Cloud v1 API supports an optional base64-encoded MD5 checksum in some requests, which can catch data integrity issues. Many languages have built-in functions to calculate the value of the content-md5 header. The following example uses Python:

$ echo "750" | python -c "import base64, hashlib; print(str(base64.b64encode(hashlib.md5(bytes(input(), encoding='utf8')).digest()), encoding='utf8'))"

sTf90fedVsft8zZf6nUg8g==

If you run into issues generating a valid content-md5 value, you might need to encode your request body in UTF-8 before computing the checksum.