Gnuplot: Ploting datasets from one file

Question

I have a file with multiple datasets that looks something like this:

#   Isochrone   Z = 0.00800     Age =   1.000e+07 yr
#   Mu  Mb  Mv  Mr  Mi  Mj  Mh  Mk  Flum
    14.982  13.538  12.020  11.076  10.083  9.024   8.401   8.172   -9.59274740
    13.741  12.481  11.074  10.195  9.366   8.378   7.710   7.510   -6.50542581
...
#   Isochrone   Z = 0.00800     Age =   1.122e+07 yr
#   Mu  Mb  Mv  Mr  Mi  Mj  Mh  Mk  Flum
    14.982  13.538  12.020  11.076  10.083  9.024   8.401   8.172   -9.59274740
    13.741  12.481  11.074  10.195  9.366   8.378   7.710   7.510   -6.50542581
...

I need to be able to use the index feature, like this:

plot 'file' index n u 1:2

so I can plot the n-th dataset, or a range of datasets (index a:b), but I get an x range is invalid error. Probably because the commented lines are not being treated as two blank lines. Could this be fixed?

Also, would it be possible to include in the title of the plot the Age value?

ignore the using parameters, I deleted a few columns, I need to plot Mb-MV on x axis, and MV on y. — Jaka
with index 0, I get a plot of the entire file. The headers aren't getting recognised as data set borders. — Jaka

Miguel Miguel · Accepted Answer · 2014-11-15T10:47:24

From the gnuplot help: "Data sets are separated by pairs of blank records". I am guessing you are not using blank lines as data set separators. If your file looks like this:

#   Isochrone   Z = 0.00800     Age =   1.000e+07 yr
#   Mu  Mb  Mv  Mr  Mi  Mj  Mh  Mk  Flum
    14.982  13.538  12.020  11.076  10.083  9.024   8.401   8.172   -9.59274740
    13.741  12.481  11.074  10.195  9.366   8.378   7.710   7.510   -6.50542581
#   Isochrone   Z = 0.00800     Age =   1.122e+07 yr
#   Mu  Mb  Mv  Mr  Mi  Mj  Mh  Mk  Flum
    14.982  13.538  12.020  11.076  10.083  9.024   8.401   8.172   -9.59274740
    13.741  12.481  11.074  10.195  9.366   8.378   7.710   7.510   -6.50542581

Then the following bash command can add the blank lines without the need to manually edit the file's content:

user@machine:~$ sed 's/#   Isochrone/\n\n#   Isochrone/g' file


#   Isochrone   Z = 0.00800     Age =   1.000e+07 yr
#   Mu  Mb  Mv  Mr  Mi  Mj  Mh  Mk  Flum
    14.982  13.538  12.020  11.076  10.083  9.024   8.401   8.172   -9.59274740
    13.741  12.481  11.074  10.195  9.366   8.378   7.710   7.510   -6.50542581


#   Isochrone   Z = 0.00800     Age =   1.122e+07 yr
#   Mu  Mb  Mv  Mr  Mi  Mj  Mh  Mk  Flum
    14.982  13.538  12.020  11.076  10.083  9.024   8.401   8.172   -9.59274740
    13.741  12.481  11.074  10.195  9.366   8.378   7.710   7.510   -6.50542581

The command above inserts two blank lines before # Isochrone every time it finds it. And now you can use the command inside gnuplot:

plot "< sed 's/#   Isochrone/\\n\\n#   Isochrone/g' file" index 0 u 1:2

Note you need to escape backslashes within gnuplot. To get the age you can use a similar approach with a system call within gnuplot:

n = 1 # First record
age_1 = system("awk '/Age/{i++}i==" . n . "{print $(NF-1); exit}' file")
print age_1
 1.000e+07

Or get all of them at the same time and store them in a string:

age = ""
do for [n = 1:2] {
age = age . " " . system("awk '/Age/{i++}i==" . n . "{print $(NF-1); exit}' file")
}
print age
 1.000e+07 1.122e+07

Now you can conveniently use the same index for index and the title:

plot for [i=1:2] "< sed 's/#   Isochrone/\\n\\n#   Isochrone/g' file" \
index (i-1) u 1:2 title "Age = " . word(age,i) . " years"

enter image description here

Are you studying magnetic flux in stars?

Gnuplot: Ploting datasets from one file

1 Answers