Right now, it’s possible to query and visualize CSV data using the CSV from URL data source. But for people who aren’t savvy with infrastructure or those needing to do a one-off report, an easier option would be most welcome. I’m envisioning an option to upload a CSV directly into Redash and treat it as a “table” on its own data source.

This could be accomplished with a link to Dropbox, Google Drive, or Box. The underlying data access logic could be shared from other current CSV-from-URL data source.

3 Likes

This definitely makes sense. The biggest blocker for this is our permissions model: currently query’s access permissions are derived from the data source it uses. If it doesn’t have a data source, then it’s in a limbo where no one can access it.

Once we implement safe parameters handling in the backend, we can move on to implement a more robust permissions model (more like Google Drive’s model). When we’re there, a user can upload a dataset (CSV or other) and assign permissions to them as they see fit.

Also, once we have this new model, we no longer will need dedicated query runners for CSV, URL, or Query Results. All this can be created by a user without the need to create the data source ahead of time, and shared as they need.

I hope the explanation makes sense, please ask to elaborate, if needed. I want to turn this into a comprehensible GitHub issue :slight_smile:

Very relevant issue:

Bumping my own thread to say that an option to upload Excel files would also be excellent, though still bound by the limitations described above.

1 Like