In Gnuplot, how can I plot the sum of two columns when I'm plotting by header name

0
votes

I have the following data file:

denst densu densd denss
3     1     1     1
4     1     1.5   1.5
5     1     2.5   1.5

I can plot, say, densu(denst) as:

plot 'file.txt' u 'denst':'densu'

Which is very convenient syntax. But if I want to plot, say, the sum of densu and densd, with respect to denst the only way I can do it is:

set key autotitle columnhead to tell gnuplot the first line is headers and not data

plot 'file.txt' u 1:($2+$3) to plot

The question is how can I do operations with column values like that, but using the name notation? The actual file is a csv with ~40 columns, and it's very tedious to manually count which column is which number so I can use the $n syntax to do math with column data.

I would want to do something like plot 'file.txt' u 1:($'densu'+$'densd'), using header name syntax analogously to how I can do it with column number syntax. Is there any way to do this?

1
plotgnuplot

1 Answers

8
votes

I've discovered a way to do it. These two commands are equivalent:

plot 'file.txt' u 1:($2+$3)
plot 'file.txt' u 1:(column(2)+column(3))

You can't do

plot 'file.txt' u 'denst':($'densu'+$'densd')

but you can do

plot 'file.txt' u 'denst':(column('densu')+column('densd'))

to the same effect.