Filtering and summation of data

[[e-mail]]

Hi,

considering data in a table where value are categorized.
I would like to be able a) to filter data for a plot by category
and b) I would like to sum up all values with same category
being able to display it in a plot with box.

Important: the solution should be platform independent;
as an example "awk" or "sed" is not available on all platforms.
Yes python would be possible but I really would prefer the
pure Gnuplot solution if possible.

Kind Regards,
Thomas

Appendix:

What I. have sofar:

<code>
set terminal pngcairo size 1400,1000 enhanced font ",20"
set output "DemoSum.png"

$DemoData << EOD
Timestamp, Category, Value
2022-05-19 05:55:00, test1, 5
2022-05-19 09:16:00, test2, 15
2022-05-19 13:56:00, test3, 10
2022-05-20 05:15:00, test1, 20
2022-05-20 10:58:00, test2, 5
2022-05-20 17:20:00, test3, 25
EOD

set datafile separator ","
set key autotitle columnhead

# a very simple plot functionality
set multiplot layout 2, 1 title "Timestamps And Sums"

# define line style(s)
set style line 1 lc rgb 'grey30' ps 0 lt 1 lw 2
set style line 2 lc rgb 'grey70' lt 1 lw 2
set style fill solid 1.0 border rgb 'grey30'

set timefmt "%Y-%m-%d %H:%M:%S"
set xdata time
set grid

set title "Data based on date and time"
plot $DemoData using 1:3 with linespoint linestyle 1

unset xdata
set title "Data based on categories"
plot $DemoData using 3:xtic(2) with boxes linestyle 1
</code>

[Jörg Buchholz]

Post by ***@googlemail.com
Hi,
considering data in a table where value are categorized.
I would like to be able a) to filter data for a plot by category
and b) I would like to sum up all values with same category
being able to display it in a plot with box.
Important: the solution should be platform independent;
as an example "awk" or "sed" is not available on all platforms.
Yes python would be possible but I really would prefer the
pure Gnuplot solution if possible.
Kind Regards,
Thomas
<code>
set terminal pngcairo size 1400,1000 enhanced font ",20"
set output "DemoSum.png"
$DemoData << EOD
Timestamp, Category, Value
2022-05-19 05:55:00, test1, 5
2022-05-19 09:16:00, test2, 15
2022-05-19 13:56:00, test3, 10
2022-05-20 05:15:00, test1, 20
2022-05-20 10:58:00, test2, 5
2022-05-20 17:20:00, test3, 25
EOD
set datafile separator ","
set key autotitle columnhead
# a very simple plot functionality
set multiplot layout 2, 1 title "Timestamps And Sums"
# define line style(s)
set style line 1 lc rgb 'grey30' ps 0 lt 1 lw 2
set style line 2 lc rgb 'grey70' lt 1 lw 2
set style fill solid 1.0 border rgb 'grey30'
set timefmt "%Y-%m-%d %H:%M:%S"
set xdata time
set grid
set title "Data based on date and time"
plot $DemoData using 1:3 with linespoint linestyle 1
unset xdata
set title "Data based on categories"
plot $DemoData using 3:xtic(2) with boxes linestyle 1
</code>

stats $DemoData u(strcol(2) eq "test1" ? $3:1/0)
sum_test1 = STATS_sum
print sum_test1

this is one way to get information for one category inside gnuplot.

Jörg

[Jörg Buchholz]

Post by ***@googlemail.com
Hi,
considering data in a table where value are categorized.
I would like to be able a) to filter data for a plot by category
and b) I would like to sum up all values with same category
being able to display it in a plot with box.
Important: the solution should be platform independent;
as an example "awk" or "sed" is not available on all platforms.
Yes python would be possible but I really would prefer the
pure Gnuplot solution if possible.
Kind Regards,
Thomas
<code>
set terminal pngcairo size 1400,1000 enhanced font ",20"
set output "DemoSum.png"
$DemoData << EOD
Timestamp, Category, Value
2022-05-19 05:55:00, test1, 5
2022-05-19 09:16:00, test2, 15
2022-05-19 13:56:00, test3, 10
2022-05-20 05:15:00, test1, 20
2022-05-20 10:58:00, test2, 5
2022-05-20 17:20:00, test3, 25
EOD
set datafile separator ","
set key autotitle columnhead
# a very simple plot functionality
set multiplot layout 2, 1 title "Timestamps And Sums"
# define line style(s)
set style line 1 lc rgb 'grey30' ps 0 lt 1 lw 2
set style line 2 lc rgb 'grey70' lt 1 lw 2
set style fill solid 1.0 border rgb 'grey30'
set timefmt "%Y-%m-%d %H:%M:%S"
set xdata time
set grid
set title "Data based on date and time"
plot $DemoData using 1:3 with linespoint linestyle 1
unset xdata
set title "Data based on categories"
plot $DemoData using 3:xtic(2) with boxes linestyle 1
</code>

Do you mean the second graph like this?

<code>
set terminal pngcairo size 1400,1000 enhanced font ",20"
set output "DemoSum.png"

$DemoData << EOD
Timestamp, Category, Value
2022-05-19 05:55:00, test1, 5
2022-05-19 09:16:00, test2, 15
2022-05-19 13:56:00, test3, 10
2022-05-20 05:15:00, test1, 20
2022-05-20 10:58:00, test2, 5
2022-05-20 17:20:00, test3, 25
EOD

set datafile separator ","
set key autotitle columnhead

# a very simple plot functionality
set multiplot layout 2, 1 title "Timestamps And Sums"

# define line style(s)
set style line 1 lc rgb 'grey30' ps 0 lt 1 lw 2
set style line 2 lc rgb 'grey70' lt 1 lw 2
set style fill solid 1.0 border rgb 'grey30'

set timefmt "%Y-%m-%d %H:%M:%S"
set xdata time
set grid

set title "Data based on date and time"
plot $DemoData using 1:3 with linespoint linestyle 1

unset xdata
set title "Data based on categories"

stats $DemoData u(strcol(2) eq "test1" ? $3:1/0)
sum_test1 = STATS_sum
stats $DemoData u(strcol(2) eq "test2" ? $3:1/0)
sum_test2 = STATS_sum
stats $DemoData u(strcol(2) eq "test3" ? $3:1/0)
sum_test3 = STATS_sum

set xtics in ('test1' 1, 'test2' 2, 'test3' 3)

set boxwidth 1 absolute
set yrange [0:*]
set xrange [0:4]
plot '+' u (1):(sum_test1) w boxes, '+' u (2):(sum_test2) w boxes, '+'
u (3):(sum_test3) w boxes
unset multiplot
set out
reset
</code>