Gnuplot: how to plot with a continuous line over multiple data blocks?

Question

I have a dat file that is split into multiple blocks, for example as the following one:

#   Time (s)    x (m)
     0.0    0.0
     1.0    1.0
     2.0    2.0
     3.0    3.0
     4.0    4.0
     5.0    5.0


     5.0   10.0
     6.0   11.0
     7.0   12.0
     8.0   13.0
     9.0   14.0
    10.0   15.0

The following minimal script:

filename = 'test.dat';

set terminal pngcairo size 960, 540 font 'Verdana, 20'
set output "test.png"
unset key
set xlabel "Time (s)"
set ylabel "x (m)"
set grid
set autoscale fix

plot filename u 1:2 w l lw 1.0 lc rgb 'black' notitle

produces this figure:

Gnuplot plots two lines, one for each block. However, since the x data are discontinuous between the two blocks, a jump shows up in the plot.

I would like to join these two lines in the output plot, without modifying the data file. Is there a way to do this?

Maybe it can be done by reading the last line of each block and the first one of the next block and plotting a line between such two points, but I am not familiar with any built-in function that can do the job.

Answer 1

The first approach which comes to my mind is plotting the data into a table. This will remove empty lines. Maybe there are better ways.

Script: (works with gnuplot>=5.0.0)

### remove empty lines in data
reset session

$Data <<EOD
#   Time (s)    x (m)
     0.0    0.0
     1.0    1.0
     2.0    2.0
     3.0    3.0
     4.0    4.0
     5.0    5.0


     5.0   10.0
     6.0   11.0
     7.0   12.0
     8.0   13.0
     9.0   14.0
    10.0   15.0
EOD

set table $Data2
    plot $Data u 1:2 w table
unset table

plot $Data2 u 1:2 w lp pt 7
### end of code

Addition: Actually, there is another approach in a single line of code (although not too obvious as well) which doesn't duplicate the data in memory (maybe only relevant for large data). You could use undefine $Data which I guess will free the memory of $Data.

The following single line code will give the same result as the table procedure above. The data is drawn as vectors from one datapoint to the next and hence ignoring empty lines as well.

plot x1=y1=NaN $Data u (x0=x1):(y0=y1):(x1=$1,x1-x0):(y1=$2,y1-y0) w vec nohead notitle

Addition 2:

As user @hanitors commented, if you have a .csv file and set datafile separator comma, the datablock $Data2 will not be plotted. You have several options:

• if your CSV has some space after the comma, i.e. 1.0, 2.0 you can keep the default separator whitespace because gnuplot will ignore the comma after the first number during number interpretation.
• if your CSV has no space space after the comma, i.e. 1.0,2.0 you can set datafile separator comma before plotting to the table and set datafile separator whitespace after. $Data2 will have space as separator and will be plotted correctly.
• if you want to continue to use comma as separator (e.g. if you want to plot other files into the same graph which have comma as separator), then you have to plot differently to the table, i.e. plot $Data u (sprintf("%g,%g",$1,$2)) w table. Then $Data2 has comma as separator as well. However, this requires gnuplot>=5.2.0.

Result: