I have the data in the following format score_count score_value 23 50 46 52 1 53 890 54 with more than a 1 million score_values, each value having its distinct count.    Given the data is in this format, how can I efficiently compute p95 or p99 of my the `score`.  I tried unrolling the table to create a single column with the value repeated multiple times, but the query does not complete at all.  Note that I'm planning to use this in a time chart command as so it has to be efficient to compute this value as well.  ... View more