Ask HN: Project to store & cache result sets?

Wasn't there a project on HN a few weeks back that was a SQL analytics tool that would store the result set of every query run? What was it called?

My actual question: I need to build a system that caches arbitrary result sets and then performs a variety of sums & groupings for click through purposes, as well as join against future result sets. I've thought of a couple different options, like storing in a separate table that's basically "Column_1", "column_2", & storing the headings somewhere else, or a JSON object of the result set. I've gone down the path of the JSON object, and it works, but for large sets converting it back and forth is turning out to be non-trivial. Has anybody solved a similar problem?