[ckan-dev] question about private/public datasets

Thu Aug 30 11:14:51 UTC 2018

Hi Adil

>Hello,
>I have played around a little with CKAN. What I would like to do is to create a dataset the is initially private accessible only to the user and stakeholders for that dataset.
>
>I thought I may be able to achieve this with groups. But, it seems not to work. What I did was to create Users A and B and register them with the same organisation O1. Then I created a group G1 and registered user A with the group. >User A created a dataset which was marked private. 
>
>However, user B was still able to see the dataset even though they were not a member of the group G1. I guess private datasets works only on the organisation? So, each dataset would need to have a separate organisation?
>
>I'm using the most recent version of CKAN.
To my knowledge, datasets are either private/public on a Organizational level or private/public on a site wide level. I'm not aware if it's possible to do it on the group level.
You could have a look at https://github.com/conwetlab/ckanext-privatedatasets, although that seems to work on a user level.

>I have another question. Does anyone use CKAN as a data archive?
Energidataservice.dk, which is Energinet's (Energinet.dk) platform for sharing public energy  data in Denmark, archives it's data using CKAN. We use the datastore as a primary datastore and generates resource files (Such as CSV, JSON etc.) dynamically from user defined filters. Our datasets are constantly being appended to. We've basically taken CKAN's data model and turned it on its head. I don't know if this constitutes a "data archive" from your point of view? You can find our public codebase at https://gitlab.com/energinet/eds 

Regards

Mads R. Havmand
IT Enterprise Information Mgmt.
+4561149549
mhv at energinet.dk