Is there a simpler way to create gnuplot filled curves for time data when data is only recorded on changes?

Question

It is easy to create a filledcurve plot when time series data is at small regular intervals, and data changes infrequently.

But what if data is recorded only on changes?

In order to make a shaded plot for when data is '1', there needs to be another point before the change in value (the ZZ.999 values):

# test shaded
#
t, v1
0, 0
9.999, 0
10, 1
14.999, 1
15, 0
29.999, 0
30, 1
44.999,1
45, 0
59.999,0
60, 1
64.999, 1
65, 0

This will give you:

This is done by:

set datafile separator "," 
set datafile commentschar '#'
set key autotitle columnhead # use the first line as title for data

fileused = 'shaded.csv'

set xtics 5
set xrange[0:90]

set term qt size 800, 400

plot fileused using 1:2 with filledcurves

Is there a way to get the same plot, without having the ZZ.999 datapoints in the dataset, so you only need the data below?

0, 0
10, 1
15, 0
30, 1
45, 0
60, 1
65, 0

Answer 1

There are several plot styles that might suit. The simplest is probably plot with fillsteps.

$DATA << EOD
0, 0
10, 1
15, 0
30, 1
45, 0
60, 1
65, 0
EOD

set datafile separator comma
set xtics 5
set xrange[0:90]
set yrange[0:1.5]

set style fill solid
plot $DATA using 1:2 with fillsteps