I have a multiple CSV files like this (just one column):
3
4
2.3
0.1
Now I want to create a gnuplot bar chart that has <filename>:<sum of the column>.
But currently I struggle with summing up a single column:
plot 'data1.txt' using 0:(sum [col = 0:MAXCOL] (col)) with linespoint;
The command you show is summing each row rather than each column.
(1) If you can transpose rows/columns in your csv file before feeding it to gnuplot, this command would produce a plot close to what you ask for. Note that MAXCOL is really the number of rows (not columns) in the original data file
set boxwidth 0.5
set style fill solid
plot 'transpose_of_original' using 0:(sum [col=0:MAXCOL] col) with boxes
(2) Alternatively you can do the summing gnuplot by first accumulating the sums and then plotting it afterward
# get number of columns
stats 'data1.txt' nooutput
NCOL = STATS_columns
array SUM[NCOL]
# get sum for each column
do for [col=1:NCOL] {
stats 'data1.txt' using col nooutput
SUM[col] = STATS_sum
}
# Now we plot the sums in a bar chart
set style fill solid
set boxwidth 0.5
set xlabel "Column"
set ylabel "Sum"
plot SUM using 1:2 with boxes
With help from @Ethan, I was able to solve my problem:
array files = ['data1.txt', 'data2.txt']
array SUM[|files|]
do for [i=1:|files|] {
stats files[i] using 1 nooutput
SUM[i] = STATS_sum
}
set style fill solid
set boxwidth 0.5
set xlabel 'File'
set ylabel 'Sum'
set yrange [0:]
plot SUM using 1:2:xticlabels(files[column(0)+1]) with boxes
data1.txt:
11
22
33
44
data2.txt:
11
2
33
4
