What do you mean by "pixels in the screen"?

A processing sketch doesn't easily have access to what is drawn outside the sketch window. If your goal is to work with the pixels being drawn by your OS and other programs, that's possibly, but is already not simple.

If you want to take 80% of the pixels from within a sketch's window, that's much easier.

What do you mean by "put them into an array"? To what end?

Do you want to pick pixels at random? Do you worry about picking the same pixel more than once?

I used the loadPixels() command, to access all the pixels in the sketch. I'm eventually trying to translate a code I did in python that was an agent based script using 3D points (in Rhino 3D) to processing using pixels instead( but I'm obviously far off from doing that right no). But to answer your questions:

1. Yes I want to pick them randomly, and I want to store the locations of each one in a list so that I could later access them individually

2. And yes, I want to pick a random 80% of the pixels without repeating.

Thanks again!

one way to do this would be to have two arrays on ints, one for X positions and the other for Y position, that are each as long as the pixels array. Assign each element it's own index (aX[i]=i), then randomize the array. Then lob off the last 20% of it with shorten() (I think).