Sort search results by number of hits per dataset, and hide datasets with no results

[rschalkrce]

Users prefer to see the dataset with the most search results on top, and then in descending order.

Datasets with no results could be grouped to enhance the readability of search results.

[rschalkrce]

Can we follow up on this issue? Now that users can select many datasets for one search, I would prefer a different sorting of the results and perhaps summarize those without results in one block.