How to automate saving plots with Shell scripting - bash

Question

I have 10 .txt files that contain my data as in output1.txt, output2.txt, output3.txt and so on. I plot this data with the following commands:
#!/bin/bash
gnuplot
set o "output1.png"
p "output1.txt" u 1:2 with lines
set o "output2.png"
p "output2.txt" u 1:2 with lines
set o "output3.png"
p "output3.txt" u 1:2 with lines
and so on.
Obviously when I have like 100 data files, this approach is not very useful. How can I write a shell script that basically automates saving my data into files with a for loop in a script? Is it even possible or should I turn somewhere else like a python script?

Answer 1

I guess you require something like this:
for each in output*.txt
do
BASE_NAME=$(basename $each .txt)
echo ${BASE_NAME}
gnuplot <<- EOF
set o "${BASE_NAME}.png"
p "$each" u 1:2 with lines
EOF
done

Answer 2

Completely untested:
#!/bin/bash
for fn in $(ls output*.txt); do
# Extract file "ID"
fid=${fn:6:(-4)}
# Build output png
out=output${fid}.png
# Echo the commands
echo set o "${out}"
echo p "${fn}" u 1:2 with lines
done
Which you'd either pipe directly into gnuplot:
./abovescript.sh | gnuplot
Or redirect into a script file, then call using gnuplot
./abovescript.sh > myscript
gnuplot myscript
Updated per Mark Setchell's suggestions:
for fn in output*.txt; do
# Extract the file "base"
fid=${fn%%.txt}
# Build output png
out=${fid}.png
# Echo the commands
echo set o \"${out}\"
echo p \"${fn}\" u 1:2 with lines
done | gnuplot
This also assumes that your "input" files are named exactly outputX.txt where X doesn't contain any whitespace.
Added by Mark Setchell... just for fun, the plots look kind of more professional if you label them a little bit, by adding the following:
...
out=...
# Echo the commands
echo set title \"${out} - plotted with gnuplot\"
echo set ylabel \"y axis\"
echo set xlabel \"x axis\"