I am working with US Census data for 2000. It is an individual-level data (14,081,466 observations) on personal characteristics, employment, wages, and occupation. For my purposes, I am interested in the "score" variable.

The variable "score" ranges from 1-5 (e.g. 1.38, 3.2 etc) and measures the level of mathematical knowledge where 1 indicates the individual has lowest knowledge and 5 indicates greatest knowledge. Moreover, it is possible that more than one individuals can have a same score value.

Now, I want to re-scale the score variable so that it equals the percentile score in 2000 (between zero and one). How should I proceed?