gnuplot-info — For questions and discussion about gnuplot

Re: [Gnuplot-info] Conditional plot with lines

[thse <t.s...@fz...>]

use "timecolumn(1)" instead of "$1"



--
View this message in context: http://gnuplot.10905.n7.nabble.com/Conditional-plot-with-lines-tp3519p17198.html
Sent from the Gnuplot - User mailing list archive at Nabble.com.

Re: [Gnuplot-info] Conditional plot with lines

[mattyd <mat...@ho...>]

I have tried this code and it works great. But I am having trouble getting
this approach to work with time-series data. I have replaced the first
column of the test.dat file with dates:

test.dat (revised)
*************
2001-1-1 2 1
2001-1-2 2 4 
2001-1-3 1 9
2001-1-4 2 16
2001-1-5 1 25
2001-1-6 2 36
*************

when I run the following lines:

set xdata time
set timefmt '%Y-%m-%d'
lastx = NaN ; lasty = NaN ; selection = 2
plot 'test.dat' using
($2==selection?($1,lastx=$1):lastx):($2==selection?($1,lasty=$3):lasty) with
linespoints

I get the following message:

Warning: empty x range [2001:2001], adjusting to [1980.99:2021.01]

and it looks like all of the x values are read as 2001.

When I do a test plot without conditions (i.e., plot 'test.dat' using 1:3
with linespoints) it looks fine, so it seems that the problem lies with the
approach of using the previous point, but I can't figure out what it is.

Can anyone please help?



--
View this message in context: http://gnuplot.10905.n7.nabble.com/Conditional-plot-with-lines-tp3519p17196.html
Sent from the Gnuplot - User mailing list archive at Nabble.com.

Re: [Gnuplot-info] X axis time scaling

[Theo <th...@gm...>]

Yes, I also noticed that the time is not the time which I see as my 
actual local time.

That's what the gnuplot 4.6 manual says:

Internally all times and dates are converted to the number of seconds 
from the year 2000.
...
The conversion to and from seconds assumes Universal Time (which is the 
same as Greenwich Standard Time). There is no provision for changing the 
time zone or for daylight savings. If all your data refer to the same 
time zone (and are all either daylight or standard) you don't need to 
worry about these things. But if the absolute time is crucial for your 
application, you'll need to convert to UT yourself.

Re: [Gnuplot-info] X axis time scaling

[LarsEr <in...@lb...>]

Thanks a lot Theo! That solved my problem. 
Still have a minor issue, it's a ruby script and gnuplot seems to handle a
Time different, don't know if it's day light saving or time zon (1 hour
differ).
/ Lars.



--
View this message in context: http://gnuplot.10905.n7.nabble.com/X-axis-time-scaling-tp17184p17194.html
Sent from the Gnuplot - User mailing list archive at Nabble.com.

[Gnuplot-info] Alien::Gnuplot module for Perl

[Craig D. <def...@bo...>]

Hi, all,

I've just released a draft Alien::Gnuplot module to CPAN.  It's intended to simplify and regularize Perl module access to Gnuplot, since there seem to be at least 6 modules supporting various forms of access.  If you are a Perl user, I would be grateful for feedback on how well (or poorly) it works on your system.  

Installing Alien::Gnuplot on a Perl system is meant to attempt to automatically install Gnuplot itself if it is not already present; and loading the module with "use Alien::Gnuplot" is meant to verify that Gnuplot is available and recent on your system.

Cheers,
Craig

Re: [Gnuplot-info] Histogram shadows

[Theo <th...@gm...>]

Hi Gerhard,

one simple way to do it:
Plot the data twice, slightly shifted and in grey.

set boxwidth 0.8 relative
xshift = 0.1    # or some other value
yshift = 0.1

plot 'Data.txt' using ($1+xshift):($2-yshift) with boxes fs solid 0.25 
lt "grey", '' using ($1):($2) with boxes fs solid

You also could create a 3D-Plot to create a shadow-like appearance. 
However, if you can plot data in 2D there is no need to plot it in 3D.

Theo.

Re: [Gnuplot-info] for loop iterating functions

[stibala <ada...@gm...>]

Thank you gentlemen, this was definitely helpful.
The definition of a special function to distinguish between all the
functions is the trick.
The code for plot is then pretty elegant.

Thank you again

Adam



--
View this message in context: http://gnuplot.10905.n7.nabble.com/for-loop-iterating-functions-tp17187p17191.html
Sent from the Gnuplot - User mailing list archive at Nabble.com.

Re: [Gnuplot-info] for loop iterating functions

[Thomas S. <t.s...@fz...>]

stibala <adam.stibal <at> gmail.com> writes:

> 
> Hello, 
> I have a matrix data set, each column contains a curve and I have a function
> fitted to each curve.
> I need to plot the original curves minus the respective fitted function and
> here comes the problem.
> My wanted code with for loop is
>  p for [i=2:11] "data.txt"  u 1:(column(i) - function>i<($1))
> so I can't imagine how to iterate through my functions numbered as the
> corresponding columns in data.
> It can be done without the loop simply
> p "data.txt" u 1:($2 - function2($1)), "data.txt" u 1:($3 - function3($1)),
> ...
> but I would like to plot it within the loop because I have a lot of data.
> All the functions are the same, but with different constants, for example
> g2(x)=exp(a2*x+b2)+c2 and g3(x)=exp(a3*x+b3)+c3.
> 
> Thank you for your help
> 
> Adam Stíbal
> 
> --

pass the loop variable as a second argument to the
function and distinguish between the different loop
steps using the ternary operator:
(here: from 2 to 5)

g(x,column)=\
 column==2?exp(a2*x+b2)+c2:\
  column==3?exp(a3*x+b3)+c3:\
   column==4?exp(a4*x+b4)+c4:\
    column==5?exp(a5*x+b5)+c5:\
     0/0

Re: [Gnuplot-info] for loop iterating functions

[Theo <th...@gm...>]

Hi Adam,

would this help to solve your problem?

a1 = 0.1	# define all your constants
a2 = 0.2
a3 = 0.3
b1 = 1.2
b2 = 2.3
b3 = 3.4
c1 = -1
c2 = -2
c3 = -3
# define a(n),b(n),c(n)-functions
a(n) = a1*(n==1) + a2*(n==2) + a3*(n==3) # and so on...
b(n) = b1*(n==1) + b2*(n==2) + b3*(n==3)
c(n) = c1*(n==1) + c2*(n==2) + c3*(n==3)
# define your function which will be subtracted
GPFUN_g = "g(n,x) = exp(a(n)*x+b(n))+c(n)"

I haven't tested it to the very end but if you using the function g(n,x) 
and something like the following...

plot for [2:11] "data.txt" using 1:(column(i) - g(i,$1))

There might be certainly a more elegant way to do it...
Anyway, let me know whether it works or not.

Theo.

[Gnuplot-info] for loop iterating functions

[stibala <ada...@gm...>]

Hello, 
I have a matrix data set, each column contains a curve and I have a function
fitted to each curve.
I need to plot the original curves minus the respective fitted function and
here comes the problem.
My wanted code with for loop is
 p for [i=2:11] "data.txt"  u 1:(column(i) - function>i<($1))
so I can't imagine how to iterate through my functions numbered as the
corresponding columns in data.
It can be done without the loop simply
p "data.txt" u 1:($2 - function2($1)), "data.txt" u 1:($3 - function3($1)),
...
but I would like to plot it within the loop because I have a lot of data.
All the functions are the same, but with different constants, for example
g2(x)=exp(a2*x+b2)+c2 and g3(x)=exp(a3*x+b3)+c3.

Thank you for your help

Adam Stíbal



--
View this message in context: http://gnuplot.10905.n7.nabble.com/for-loop-iterating-functions-tp17187.html
Sent from the Gnuplot - User mailing list archive at Nabble.com.

Re: [Gnuplot-info] X axis time scaling

[Theo <th...@gm...>]

Hi Lars,

does the following work for you?

set xdata time
set format x "%H:%M"
set xrange [ time(0) - 86400 : time(0) ]    # 86400 sec = 1 day
set label 1 strftime("%T",time(0)) at screen 0.5,0.5   # show the 
current time

Theo.

[Gnuplot-info] X axis time scaling

[LarsEr <in...@lb...>]

I want to create a plot that have an x range over the last 24 hours with the
current time as last position, is that possible? E.g start at 9.00 and stop
at 8.59. 

Thanks

Lars



--
View this message in context: http://gnuplot.10905.n7.nabble.com/X-axis-time-scaling-tp17184.html
Sent from the Gnuplot - User mailing list archive at Nabble.com.

Re: [Gnuplot-info] Filtering + histogram

[<kev...@fr...>]

"Thomas Sefzick" <t DOT sefzick AT fz-juelich DE de writes:

>> ...
> using the 'smooth frequency' option of the 'plot' command
> you could set the y-value to 0.0 for unwanted data instead
> of making data points undefined:
>
> plot 'gp' u (hist($2,1)):(($1==3)?1.0:0.0) sm f w boxes

Works -- thanks!

Ugh, why didn't I think of that?  (Don't answer that! :-)

I had originally tried qualifying the output rather than the input, i.e., your solution, but with 1/0 instead of 0.0 -- with the same incorrect results as what I originally posted.  Still, should either one work?  It seems like a minor bug in gnuplot (though at least one with a workaround).

------------------------------------------------------------------------------
Symantec Endpoint Protection 12 positioned as A LEADER in The Forrester  
Wave(TM): Endpoint Security, Q1 2013 and "remains a good choice" in the  
endpoint security space. For insight on selecting the right partner to 
tackle endpoint security challenges, access the full report. 
http://p.sf.net/sfu/symantec-dev2dev
_______________________________________________
gnuplot-info mailing list
gnu...@li...
https://lists.sourceforge.net/lists/listinfo/gnuplot-info

Re: [Gnuplot-info] Problem with Gnuplot command line limits

[Ethan A M. <eam...@gm...>]

On Saturday, 09 March 2013, Ludovic GIRAUD wrote:
> Hi fellow gnuplot users,
> 
> >
> >    I made a C++ program that uses gnuplot as an interface, everything
> > works fine
> > unless the command given to gnuplot exceeds a limited size (20 I think):
> >
> > This is how the command is constructed:
> >
> > string command="plot \"wallet.txt\" using 1 w l";
> > for (int i=2;i<vectwalletvect.size();i++)
> > {
> >  command.append(", \"wallet.txt\" using "+inttostring(i)+" w l");
> > }
> > g11.cmd(command);
> >
> > Works well except when i is above 20...
> >
> > Please help!

I can't help you with the C++ aspect, but placing such a command
in a file and then feeding it to gnuplot
    gnuplot < longcommand.gp
or  cat longcommand.gp | gnuplot
works fine here.  I tried up to 80 copies in the above format
before I got bored, for a total line length of >2K characters.  
Perhaps the limitation you run into has to do with the pipe 
implementation in g11.cmd (which you don't show)?

For what it's worth, the same effect could be produced by 
a single very short gnuplot command:
  plot for [i=1:20] "wallet.txt" using i w l

	Ethan

Re: [Gnuplot-info] Problem with Gnuplot command line limits

[Ludovic G. <b00...@es...>]

Hi fellow gnuplot users,

>
>    I made a C++ program that uses gnuplot as an interface, everything
> works fine
> unless the command given to gnuplot exceeds a limited size (20 I think):
>
> This is how the command is constructed:
>
> string command="plot \"wallet.txt\" using 1 w l";
> for (int i=2;i<vectwalletvect.size();i++)
> {
>  command.append(", \"wallet.txt\" using "+inttostring(i)+" w l");
> }
> g11.cmd(command);
>
> Works well except when i is above 20...
>
> Please help!
>
> LG
>

Re: [Gnuplot-info] Filtering + histogram

[Thomas S. <t.s...@fz...>]

 <kevintheobald <at> frontier.com> writes:

> 
> Given the following data (simplified example):
> 
> 3 2
> 2 3
> 3 2
> 3 2
> 
> I can plot a vertical-bar histogram with 1-unit buckets with the data in
column 2 as follows:
> 
> hist(x,width)=width*floor(x/width)+width/2.0
> set boxwidth 1
> plot 'gp' u (hist($2,1)):(1.0) sm f w boxes
> 
> producing the expected result, with one bar of height 3 from 2-3 and one bar
of height 1 from 3-4.
> 
> Now I want to show data only from rows with 3 in column 1.  Most info I've
seen on this recommends using ?: to turn
> unwanted rows into garbage, e.g., 1/0:
> 
> plot 'gp' u (hist((($1==3)?$2:(1/0)),1)):(1.0) sm f w boxes
> 
> Unfortunately, I don't get what I expect (a single bar of height 3 from 2-3).
 Instead, I get a bar of height 2
> with a horizontal line in the middle at y=1.  I think it's actually
superimposing two bars, one of height 1
> and one of height 2, because if I add another row '3 2' at the end, the height
goes up to 3 (instead of 4) and the
> line is still at y=1.  It's as if the "bad" data in row 2 "breaks up" the data
and plot doesn't realize row 1 and
> rows 3-4 really belong to the same bucket, and it draws two bars, one in front
of the other.
> 
> Is there another way to do this?  (Would prefer within gnuplot, rather than
"filter the data before going
> into gnuplot.)  Or is this a known problem recently fixed?  (I'm using 4.6.0.)
> 

using the 'smooth frequency' option of the 'plot' command
you could set the y-value to 0.0 for unwanted data instead
of making data points undefined:

plot 'gp' u (hist($2,1)):(($1==3)?1.0:0.0) sm f w boxes

[Gnuplot-info] Dashed or dotted borders of objects

[Theo <th...@gm...>]

Hi all,

so far without success, I tried to create an object 
(rectangle,circle,ellipse) with dashed or dotted border.
I am using 'term pngcairo dashed' and arrows are shown nicely with the 
different linestyles.

Something like the following does not work:
set object rect at 0.5,0.5 size 0.5,0.5 fs empty border ls 2
Gnuplot doesn't like 'ls 2' after border.

With the following
set object rect at 0.5,0.5 size 0.5,0.5 fs empty border lt 0
I expected to get a dotted border, but no, it's solid!
However, if I draw 'plot x with lines lt 0' I get a dotted line.
Why not with the border of a rectangle, circle or ellipse?

A workaraound for rectangles could be the drawing of four dashed/dotted 
arrows without heads.
What about cirles and ellipses?
Any ideas?

Theo.

[Gnuplot-info] Filtering + histogram

[<kev...@fr...>]

Given the following data (simplified example):

3 2
2 3
3 2
3 2

I can plot a vertical-bar histogram with 1-unit buckets with the data in column 2 as follows:

hist(x,width)=width*floor(x/width)+width/2.0
set boxwidth 1
plot 'gp' u (hist($2,1)):(1.0) sm f w boxes

producing the expected result, with one bar of height 3 from 2-3 and one bar of height 1 from 3-4.

Now I want to show data only from rows with 3 in column 1.  Most info I've seen on this recommends using ?: to turn unwanted rows into garbage, e.g., 1/0:

plot 'gp' u (hist((($1==3)?$2:(1/0)),1)):(1.0) sm f w boxes

Unfortunately, I don't get what I expect (a single bar of height 3 from 2-3).  Instead, I get a bar of height 2 with a horizontal line in the middle at y=1.  I think it's actually superimposing two bars, one of height 1 and one of height 2, because if I add another row '3 2' at the end, the height goes up to 3 (instead of 4) and the line is still at y=1.  It's as if the "bad" data in row 2 "breaks up" the data and plot doesn't realize row 1 and rows 3-4 really belong to the same bucket, and it draws two bars, one in front of the other.

Is there another way to do this?  (Would prefer within gnuplot, rather than "filter the data before going into gnuplot.)  Or is this a known problem recently fixed?  (I'm using 4.6.0.)

[Gnuplot-info] Binary input plot performance

[Viallard A. <via...@sy...>]

Hi there,

I see a weird thing with gnuplot.

I make plot of datas given by a velocity sensor. For each sample, there 
are 3 values of 8 Bytes.

Earlier, the plot was made using a homemade utility to convert binary 
values to ascii. These ascii datas were put into a file which was used 
by gnuplot. It was the easier solution for me.

As this process take some time, I've try to do the same plot directly 
with the binary file. I guessed it will be faster to make plot. And one 
day (actually, it's today), after a lot of manual reading and some weird 
troubles, I've successfuly done this.

After a party for celebrate my little victory, I do a time performance 
check and I'm disapointed about the results. Here the output 
(plot-txt.sh do the first solution, plot.sh do the binary solution) :

--------------------------------------------------
# time /usr/bin/plot-txt.sh /tmp/event-dclPWz
Generating plot '/tmp/test-txt.png' done !
real	0m 36.99s
user	0m 13.15s
sys	0m 2.95s
# time /usr/bin/plot.sh /tmp/event-dclPWz
Generating plot '/tmp/test.png' done
real	0m 44.37s
user	0m 12.19s
sys	0m 11.09s
--------------------------------------------------

Binary solution is slower than ascii ! It's sad.

How this can be possible ?

Here my "binary plot" script :

--------------------------------------------------
set terminal png linewidth 1 interlace size 800,600
set output "/tmp/test.png"
set title ' signals / time'
set xlabel ' Time (s) '
set ylabel ' Signals (mm/s) '
set xdata time
#set timefmt "%S"
set label "Peak : X=14.6 Y=10.4 Z=0.248" at graph 0.05,0.05
set grid
set style data lines
inscale(r)=(r*100)
genx(i)=(i*(1.0/1000))
set xrange [0:(2620*1.0/1000)]

plot \
    "/tmp/event-dclPWz" binary skip=512 format="%lf%*lf%*lf" using 
(genx($0)):(inscale($1)) title "X" with lines,\
    "" binary skip=512 format="%*lf%lf%*lf" using 
(genx($0)):(inscale($1)) title "Y" with lines,\
    "" binary skip=512 format="%*lf%*lf%lf" using 
(genx($0)):(inscale($1)) title "Z" with lines
--------------------------------------------------

There a way to improve efficiency of this script ? Why does binary plot 
take more time ?

Anthony V.

-- 
.-------------------.-----------------------.-------------------.
| Viallard Anthony  | Syscom Instruments SA | Embedded System   |
| +41 024 455 24 82 | Rue de l'industrie 21 | Software Designer |
| ~~~~~~~~~~~~~~~~~ | 1450 Sainte-Croix     | ~~~~~~~~~~~~~~~~~ |
'-------------------^-----------------------^-------------------^

Re: [Gnuplot-info] log y axis scales

[Theo <th...@gm...>]

Hi Al,

have you tried the following?

set format y "%.1l*10^{%L}"

Theo.

Am 06.03.2013 03:40, schrieb Al Conle:
> Hi,
>     While using   gnuplot version 4.6 pl 1  I found that the scale
> annotations
> for a logscale y   axis  were  remaining in fixed decimal format rather
> than
> going to exponential representation.  For numbers less than   0.0000001
>
> all the annotations are  0.0000000  .   If one plots a linear scale the
> annotations
> are correct.
>     The graph with the flaw is attached.
>
> Al
>
>

[Gnuplot-info] log y axis scales

[Al C. <fa...@gm...>]

Hi,
   While using   gnuplot version 4.6 pl 1  I found that the scale
annotations
for a logscale y   axis  were  remaining in fixed decimal format rather
than
going to exponential representation.  For numbers less than   0.0000001

all the annotations are  0.0000000  .   If one plots a linear scale the
annotations
are correct.
   The graph with the flaw is attached.

Al

Re: [Gnuplot-info] A weird case with implicit columns

[Craig D. <def...@bo...>]

Thanks!

On Mar 4, 2013, at 8:17 PM, sfeam <sf...@us...> wrote:

> On Thursday, February 28, 2013 09:59:53 PM Craig DeForest wrote:
>> Some 'with' modifiers include implicit dimensions in ASCII mode, others
>> don't.
>> 
>> 
>> 
>> For example:
>> 
>> 	plot '-' using 1 with lines
>> 
>> works, while
>> 
>> 	plot '-' using 1:2 with yerrorbars
>> 
>> does not.
> 
> I've added that combination (implicit x with yerrorbars) for
> 4.7 and 4.6.2
> 
> 	Ethan
> 
> 

[Gnuplot-info] iOS / iPhone compilation

[Casey B. <cas...@gm...>]

Hi,

Has anyone here compile gnuplot for iOS?

Any advice towards that end would be greatly appreciated.

Thanks,
Caset


-- 
Casey James Basichis
Composer - Adventure Time - Cartoon Network
http://www.caseyjamesbasichis.com
cas...@gm...
310.387.7540

Re: [Gnuplot-info] splot errorbars in fence plot

[atoms.h <tho...@gm...>]

Thanks a lot Ethan, that worked beautifully!

(As I understand it, the vectors style plots the delta from the previous
position of the coordinate. So I had to fill in two lines for both
directions of the error bars, but that's no problem at all.)

About the pdf terminal: yes, in that example the dl was too small, but that
was only where I stopped trying. Previously I made a 2D (using 'plot') plot
of one of the panels with the same terminal setup and got actual dashed
lines, but when using splot they disappear....that's why I took a detour
with  the eps enhanced ps terminal and execute !epstopdf at the end of the
script. That's how I got the pdf with dashed in the example (same size, same
dl settings)...

again, thanks a lot!



--
View this message in context: http://gnuplot.10905.n7.nabble.com/splot-errorbars-in-fence-plot-tp17125p17143.html
Sent from the Gnuplot - User mailing list archive at Nabble.com.

Re: [Gnuplot-info] wxt vs. png output

[Theo <th...@gm...>]

Thanks Ethan,

that's it.
pngcairo does it as expected!