Skip to main content

Register Datasets

Register your datasets with the NDE Dataset Register.

What happens after you register​

You register a URL that points to one or more dataset descriptions, not the descriptions themselves. The Register fetches that URL, validates what it finds, and only then stores it.

  • If your description is valid, the URL is registered and you are done. The Register re-crawls it daily to pick up your changes, so you do not have to register again when a description changes.
  • If your description is invalid, the registration is refused and nothing is kept: the Register will not come back to your URL by itself. Once you have fixed the problem, you have to register the URL again.

Registering the same URL twice is harmless, so when in doubt, register again.

caution

A description can be refused for a reason that is not in the description itself. If it declares a distribution (a data dump or a SPARQL endpoint) that the Register cannot reach at that moment, the description is refused as invalid even if the metadata is perfectly fine. If you cannot offer a service reliably, do not declare it: removing an unreliable SPARQL endpoint from the description is enough to make it registrable.

Note the asymmetry: an unreachable distribution blocks a new registration, but only raises a warning on one that already exists. See Validation.

Check before you register​

Because a refused submission leaves no trace, check the description before registering it, with the validation web service or the validation endpoints. They apply the same rules as registration, including the distribution checks, so they tell you in advance whether registration would succeed.

If you publish through a collection management system, it should be doing this for you and showing you the result: that is required of it. If it lets you believe a description was registered when it was refused, report it to your supplier.