gnuplot linear fit within for loop

Question

I have a script that correctly plots 100 separate data files. Now I want to do a linear fit of each plot within a for loop. However, my script crashes when

Answer 1

fit doesn't support iterations. Here is how you can do it, but it needs some fiddling ;)

# define the functions depending on the current number
fstr(N) = sprintf("f%d(x) = a%d*x + b%d", N, N, N)

# The fitting string for a specific file and the related function
fitstr(N) = sprintf("fit f%d(x) 'file%d.dat' via a%d,b%d", N, N, N, N)

n = 2

# Do all the fits
do for [i=1:n] {
    eval(fstr(i))
    eval(fitstr(i))
}

# construct the complete plotting string
plotstr = "plot "
do for [i=1:n] {
    plotstr = plotstr . sprintf("f%d(x), 'file%d.dat'%s ", i, i, (i == n) ? "" : ", ")
}

eval(plotstr)

This works fine for me, if I use the following two test files:

File file1.dat:

1 1
2 2.1
3 3

and file2.dat:

1 1.5
2 2.7
3 4

The result with 4.6.5 is:

To have the actual result of the fitting displayed in the key you must construct the plotting string plotstr as follows:

plotstr = "plot "
do for [i=1:n] {
    t = sprintf("f%d(x) = %.2f*x + %.2f", i, value(sprintf('a%d', i)), value(sprintf('b%d', i)))
    plotstr = plotstr . sprintf("f%d(x) lt %d t '%s', ", i, i, t).\
                        sprintf(" 'file%d.dat' lt %d %s ", i, i, (i == n) ? "" : ", ")
}

with the result