Gnuplot Cumulative Column Question

Question

I have some data.

#Time  Distance
 1   3
 2   5
 4   9
 8  11
12  17
14  20
16  34
20  40

I want to plot the cumulative distance wrt time in gnuplot ... (it should be easy) but I do not know how.

x

Answer 1

For anyone still looking for this sort of thing, If your gnuplot version is 4.4 or newer, you can do the following:

a=0
#gnuplot 4.4+ functions are now defined as:  
#func(variable1,variable2...)=(statement1,statement2,...,return value)
cumulative_sum(x)=(a=a+x,a)
plot "test.dat" using 1:(cumulative_sum($2))

Answer 2

If your data is in the file datafile, you can do a cumulative plot like so:

$ gnuplot
gnuplot> plot "datafile" smooth cumulative

Answer 3

Assuming your data is in a file "test.txt", how about:

plot "<awk '{i=i+$2; print $1,i}' test.txt" with lines

Answer 4

The same can be achieved by the "cumulative" variant of the "smooth" option (just enter help smooth in gnuplot).

Answer 5

I have a little experience with Gnuplot and I just pored over the documentation some. Unfortunately, I wasn't able to find a solution for generating a cumulative sum as you're plotting.

I think what you'll need to do is to massage your data with another program before letting Gnuplot at it. awk is one program that comes to mind, it's practically built for fiddling with columnar data. You can integrate this process into the plotting process by following these instructions.

Answer 6

This isn't what the question is looking for but "gnuplot cumulative distribution function" always leads me here and none of the answers are quite right unless you already have the data gnuplot expects.

But suppose you wanted a cumulative distribution function of just distances and only have the raw distances. There is a little trick with which gnuplot can calculate the y values:

test.dat:

#Time  Distance
 1   3
 2   5
 4   9
 8  11
12  17
14  20
16  34
20  40

plot.gpi:

datafile = test.dat
stats datafile
plot datafile using 2:(1./STATS_records) smooth cumulative title "distance"

It is not as flexible as having all the data, though.