How to use a column in gnuplot to separate plots

Question

I have the following data to draw the CPU usage of running tasks at specific time stamps.

cpu timestamp load taskId
0 0 1.0 0
1 0 1.0 0
0 0 0.5 1
1 0 0.5 1
0 65 0.5 0
1 65 0.5 0
0 20094 0.0 0
1 20094 0.0 0
0 40042 0.0 1
1 40042 0.0 1

This is what I use to create the plot using gnuplot:

set key autotitle columnheader
set xtics rotate
#set xrange [-0.5:100000]
set ylabel "CPU Load"
set xlabel "Time"
set key right
set grid y x
#set yrange [-0.05:1.01]
plot 'test.dat' using (column(2)):3 with linespoints
pause -1 "Hit return to continue"

The result is:

However, I would like to plot two lines one each for a task ID, using the fourth column. The result should be like:

How can I fix my graph? If I can also plot two graphs using the first column, that would be even better.

Answer 1

set datafile missing NaN
set offset graph 0.05, graph 0.05
set ylabel "CPU Load"
set xlabel "Time"
set key outside right
set grid y x
set xtics rotate
set style data linespoints
set autoscale noextend
set border back

set multiplot layout 2,1
do for [i=0:1] {
    set title sprintf("CPU %d", i)
    plot 'data' using (column(1) == i ? column(2) : NaN) : 3 title "Task 0" pt 7 lw 2, \
         'data' using (column(1) == i ? column(2) : NaN) : 4 title "Task 1" pt 7 lw 2
}
unset multiplot