Using bash script variable in gnuplot plot command

Question

I want to plot a set of data stored in different files:

file.txt_100K
file.txt_200K
file.txt_300K

I came up with the following bash script:

  1 #!/bin/bash
  2 
  3 filename=file.txt_
  4 
  5 gnuplot <<- EOF
  6     set term png
  7     set output "plot.png"
  8     plot '"$filename"100K' w lp,\
  9                 '"$filename"200K' w lp,\
 10                 '"$filename"300K' w lp
 11 EOF

The following is the output after running the script:

         line 0: warning: Cannot find or open file ""file.txt_"100K"
     line 0: warning: Cannot find or open file ""file.txt_"200K"
     line 0: warning: Cannot find or open file ""file.txt_"300K"
     line 0: No data in plot

I was hoping that in lines 8, 9 and 10 of the bash script the following would be the case:

'"$filename"100K' == 'file.txt_100K'
'"$filename"200K' == 'file.txt_200K'
'"$filename"300K' == 'file.txt_300K'

What change should I make to the lines 8, 9, 10 to get the above desired result ? Thanks in advance.

Answer 1

From the man page, bash here-documents only do

parameter expansion, command substitution, and arithmetic expansion,

so it doesn't do quote removal. Use lines like

plot '${filename}100K' w lp,\