[Gnuplot-info] “block“ delimiter

[Elias A. <eli...@gm...>]

I have a file that looks like this:

   bandindex:           1
   -0.21165  -0.36658   0.28144   0.00000 -24.30536
   -0.22193  -0.36065   0.28144   0.01188 -24.30546
   -0.23222  -0.35471   0.28144   0.02375 -24.30536
…
   -0.21508  -0.35043  -0.12440   2.31382 -24.28355
   -0.21336  -0.35851  -0.13256   2.32543 -24.28356
   -0.21165  -0.36658  -0.14072   2.33703 -24.28353
   bandindex:           2
   -0.21165  -0.36658   0.28144   0.00000 -24.30536
   -0.22193  -0.36065   0.28144   0.01188 -24.30546
   -0.23222  -0.35471   0.28144   0.02375 -24.30536
…

Columns 4 and 5 hold the data I want to plot.  Column 4 increases 
monotonically, then jumps back to 0 and starts increasing again.  These 
“logical blocks” are delimited by the “bandindex“ lines.

If I plot this as “p 'file' u 4:5 w p”, everything works fine.  But what 
I would like to do is plot it with lines; then I get diagonal lines 
through my plot whenever column 4 jumps.  How can I avoid these lines in 
a comfortable way?

Of course I can do “p "<perl -pe 's/band.*//' file" u 4:5 w l”, then the 
“logical blocks” become “gnuplot blocks” suppressing the diagonal lines. 
  But I do this a lot interactively so that would be a pain to type.  It 
is also not practical for me to change the format of the file.

One way that I could imagine this to work is if I could tell gnuplot 
that lines "*band*" should delimit blocks (or lines containing 
non-numeric data, say).

Another way would be if I could define an “input filter” to be applied 
to all data files.

Is anything like that possible in gnuplot?

	Elias