I woulnd't bother with SAS at all. Why waste time on closed-source, proprietary crap when there are other alternatives available? Oh yeah, it's popular, but so are other things, and we're seeing firms looking to ditch SAS in favor of Python, R, Octave, etc. these days.
If you want three languages / platforms for data science, a fine set would be Python, R, and (Octave|Scala|Clojure|Julia).