Support rollup / filtering on simple (json, csv, flat) tabular data

Awesome thing: re:dash supports URL loading: http://docs.redash.io/en/latest/datasources.html#url

Even Awesomer if it supported loading rows and doing rollups / splits / filters.

Prior art:

• d3.csv
• Druid (http://druid.io/docs/latest/design/index.html)

It’s a big ask, but can be a medium ask if you make strong limitations

Thanks for any thoughts on this !

GL

Being able to filter or have more flexible URL loader is definitely on the radar. There are two approaches I’m considering:

1. Use some sort of JSON query language (there are several). When defining a “query” that loads data from a URL, you will provide the URL and optional JSON query. This will allow for filtering, but not roll ups.
2. Allow running SQL queries on top of the loaded data (or any other query result in Redash).

And there is of course the option to support both While #2 might sound complex, it’s not that hard to implement using approach similar to q – load the query result (or data from URL) into a temporary SQLite database, run the query and provide the result.