Reputation: 980
I want to use gnuplot to plot several lines from a csv file which has a not fixed number of header lines and columns. The file might look like this:
HEADERLINE1 = TEXTTEXTTEXT
HEADERLINE2 = TEXTTEXTTEXT
HEADERLINE3 = 1248243
HEADERLINE4 = 329340
COLUMN_X COLUMN_1 COLUMN_2
1 36.9194 30.0000
2 19.0977 20.0000
3 12.9250 10.0000
4 10.7134 0.0000
Normally without the header lines I would use something like this:
stats 'datafile.csv' nooutput
set key autotitle columnheader
plot for [t=2:STATS_columns] \
datafile \
using 1:t \
with linespoints
However with the header lines those are used by the autotitle command. Is there a way to tell autotitle to pick the columnheads from the actual columns / second block / ignore the header?
Thanks
Upvotes: 0
Views: 1448
Reputation: 7659
Provided that there are always at least two blank lines between the header lines and the data, you can use index
. stats
gives you some warnings but the result is as desired:
stats "so.dat" nooutput
set key autotitle columnheader
plot for [t=2:STATS_columns] \
"so.dat" \
index 1\
using 1:t \
with linespoints
yields
Upvotes: 2