gnuplot-info — For questions and discussion about gnuplot

Re: [Gnuplot-info] Compute values with 'every'

[Alexander S. <ma...@st...>]

Hey,

what about:

y=NaN
plot    file1 every 4::0 using 1 title "25" ls 1, \
        file1 every 4::1 using 1 title "50" ls 3, \
        file1 every 4::2 using 1 title "75" ls 5, \
        file1 every 4::3 using 1 title "100" ls 7, \
        file1 u 0:(int($0)%4==0 ? y=$1 : int($0)%4==2 ? $1*y : NaN) t
"50*100"

Best,
Alex

On Tue, 1 Dec 2020 at 15:18, Robert von Knobloch <bo...@en...> wrote:

> Hi,
> I am plotting groups of 4 lines of data.
> I can plot the first datum of each line using:
>
> plot    file1 every 4::0 using 1 title "25" ls 1, \
>         file1 every 4::1 using 1 title "50" ls 3, \
>         file1 every 4::2 using 1 title "75" ls 5, \
>         file1 every 4::3 using 1 title "100" ls 7
>
> Which works fine for the whole file of 50 or so entries
>
> What I would like to be able to do is to add a computed line e.g. the
> result of datum#1 in the first of the 4 block multiplied by datum#1 of
> the third block (for the whole file).
>
> Is there a method of doing this?
>
> Regards,
> Robert von Knobloch
>
>
> _______________________________________________
> gnuplot-info mailing list
> gnu...@li...
> Membership management via:
> https://lists.sourceforge.net/lists/listinfo/gnuplot-info
>

[Gnuplot-info] Compute values with 'every'

[Robert v. K. <bo...@en...>]

Hi,
I am plotting groups of 4 lines of data.
I can plot the first datum of each line using:

plot	file1 every 4::0 using 1 title "25" ls 1, \
	file1 every 4::1 using 1 title "50" ls 3, \
	file1 every 4::2 using 1 title "75" ls 5, \
	file1 every 4::3 using 1 title "100" ls 7

Which works fine for the whole file of 50 or so entries

What I would like to be able to do is to add a computed line e.g. the 
result of datum#1 in the first of the 4 block multiplied by datum#1 of 
the third block (for the whole file).

Is there a method of doing this?

Regards,
Robert von Knobloch

[Gnuplot-info] About using gnuplot in a non-passive mode

[edulsa <ed...@qu...>]

Dear

I have used Gnuplot (win and BSD versions) in my  Physical Chemistry 
classes since January. I have learned a lot about it and I got 
encouraged my students to use this beautiful free program.
However, I have a problem with my interface that I couldn´t solve even 
reading tutorials and manuals: I start a document, inserting constants, 
writing functions and plotting graphs without problems. I save the 
document (.plt document type) and close Gnuplot. But, when I reopened 
the same document (loading in a clean start), only the last plot save is 
showed. I would like to reopen the document to show to my students line 
by line, how to insert constants, define functions and not get only the 
last graph. Of course, I  always can edit it using vim or another editor 
but I think that it should be a more cleaver way to do this directly in 
Gnuplot plot.

My terminal is wxt. Is this caused by the terminal type that I choose?
Please, could you give suggestions?

I wish to thank you in advance for your attention

My best regards

Eduardo

-- 
Eduardo Lemos de Sa
Professor Titular
Dep. Quimica da Universidade Federal do Paraná
fone: +55(41)3361-3300
fax:   +55(41)3361-3186
Voip Number 10531185

[Gnuplot-info] get number of columns for each index

[Nunzio L. <nun...@gm...>]

Hi all,

I am using gnuplot 5.5
I have a multi-data-set file with a different number of columns for each
index.
How can I get the number of columns for each index and store it in a
variable to perform iterations?

Best,

NL

Re: [Gnuplot-info] Different font size rendering on different machines with pdf terminal

[Alexander S. <ma...@st...>]

Hey,

there might be also the fontscale parameter. But not sure why this should
be different on different machines (if you use the same initialisation)
A work around could be to read out the computer name and depending on this
you set the terminal automatically to font size 6 or 12,
e.g.:
if(system("hostname") eq "PC1") { font = 12}
else { font = 6}
set terminal pdf font "Times-New-Roman,".font

Best,
Alex

On Sat, 17 Oct 2020 at 17:11, Helmut Haberzettl <he...@ha...>
wrote:

> Hi,
>
> I use gnuplot 5.2 (patchlevel 8) on various machines running Windows 10
> Pro.
>
> When creating plots using the pdf terminal, the font sizes come out
> differently on different machines. Specifically, while
> set terminal pdf font "Times-New-Roman,12"
> works as expected on one machine, I need to use
> set terminal pdf font "Times-New-Roman,6"
> on another machine to produce approximately (but not identically) the
> same output on another machine. The plot files are identical otherwise.
>
> Are there some additional settings I'm unaware of that prevent this from
> happening? This is quite annoying because I frequently work on the same
> gnuplot file switching between machines (desktop, laptop) and any help
> in this regard would be greatly appreciated.
>
> Thank you,
> --Helmut Haberzettl
>
>
> _______________________________________________
> gnuplot-info mailing list
> gnu...@li...
> Membership management via:
> https://lists.sourceforge.net/lists/listinfo/gnuplot-info
>

Re: [Gnuplot-info] Point outside of plot field in wrong place using linespoints

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

On Friday, 23 October 2020 09:30:42 PDT Al Conle wrote:
> Hi,
> When I have a set of points,  one of which is outside of the yrange or
> xrange,  and I plot them all with   linespoints,  the line that connects
> the last plotted point to the point outside of the field can be
> significantly
> wrong.  It is as if the outside point is not in the same co-ordinate system.

Fixed in release 5.2.5 (Oct 2018).

	Ethan


> 
> E.g. plot script------------------
> set term qt enhanced font "arial,16"
> set key noenhanced
> set colors default
> set grid
> 
>  set logscale xy
>  set xlabel "2Nf,  Reversals to Failure"
>  set ylabel "Strain Ampl."
>  set yrange [0.0001:1]
>  set xrange [1:1E+08]
>  set pointsize 2
>  plot \
>  "merged3_SAE950X_fitted.html" u 2:1 w lp pt  4 lc -1
> 
> #E.g. ------------------example data file -------------------
> 1.01           1  751    0  1.006    203000 12409  #Fitted_point  monotonic
> 0.01914      640  421.6  0  0.01707  203000  1280  #Fitted_point manually
> 0.00968     2300  393.6  0  0.00774  203000   879  #Fitted_point  manually
> 0.00739     4000  382.1  0  0.00551  203000   757  #Fitted_point  manually
> 0.00532     8000  367.4  0  0.00351  203000   630  #Fitted_point  manually
> 0.00424    15000  356.6  0  0.00249  203000   554  #Fitted_point  manually
> 0.00346    24000  346    0  0.00176  203000   493  #Fitted_point  manually
> 0.00273    50000  332.1  0  0.0011   203000   429  #Fitted_point  manually
> 0.00234    80000  321.5  0  0.00076  203000   391  #Fitted_point  manually
> 0.00204   164000  310.8  0  0.00051  203000   359  #Fitted_point  manually
> 0.00175   400000  296.1  0  0.00029  203000   324  #Fitted_point  manually
> 0.00158  1000000  284.3  0  0.00018  203000   302  #Fitted_point  manually
> 0.00145  2400000  271.9  0  0.00011  203000   283  #Fitted_point  manually
> 0.00131 15000000  254.6  0  0.00005  203000   260  #Fitted_point  manually
> 0.00131 20000000  254.6  0  0.00005  203000   260  #Fitted_point  manually
> 
> 
> Example plot is attached
> 
> gnuplot version is
> Version 5.2 patchlevel 2    last modified 2017-11-15
> Running opensuse 15.1
> 
> best regards,
> Al
> 
> _______________________________________________
> gnuplot-info mailing list
> gnu...@li...
> Membership management via: https://lists.sourceforge.net/lists/listinfo/gnuplot-info
> 

[Gnuplot-info] Change color of mouse zoom, format, ruler, etc?

[mark_at_yahoo <mar...@ya...>]

Can the color of the interactive mouse feedback in the graphics terminal 
window be changed?

This of course works with problems:
gnuplot> plot sin(x) with lines linecolor "black"

But with the following the interactive mouse elements are still black 
and thus invisible on the black background:
gnuplot> set object 1 rectangle from screen 0,0 to screen 1,1 fillcolor 
rgb "black" behind
gnuplot> plot sin(x) with lines linecolor "white"

Apologies if this has been answered before -- my searches turned up 
nothing. FWIW, I'm using:
$ gnuplot
	G N U P L O T
	Version 5.2 patchlevel 2    last modified 2017-11-15
	Copyright (C) 1986-1993, 1998, 2004, 2007-2017
	Thomas Williams, Colin Kelley and many others
	gnuplot home:     http://www.gnuplot.info
	faq, bugs, etc:   type "help FAQ"
	immediate help:   type "help"  (plot window: hit 'h')
	Type `load "all.dem"` to display a large number of examples.
	They are located at /usr/share/doc/packages/gnuplot/demo/*
Terminal type is now 'qt'

If not in 5.2, is changing the colors implemented in later versions?

Thanks for any info.

-- 
MARK
mar...@ya...

[Gnuplot-info] Point outside of plot field in wrong place using linespoints

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

Hi,
When I have a set of points,  one of which is outside of the yrange or
xrange,  and I plot them all with   linespoints,  the line that connects
the last plotted point to the point outside of the field can be
significantly
wrong.  It is as if the outside point is not in the same co-ordinate system.

E.g. plot script------------------
set term qt enhanced font "arial,16"
set key noenhanced
set colors default
set grid

 set logscale xy
 set xlabel "2Nf,  Reversals to Failure"
 set ylabel "Strain Ampl."
 set yrange [0.0001:1]
 set xrange [1:1E+08]
 set pointsize 2
 plot \
 "merged3_SAE950X_fitted.html" u 2:1 w lp pt  4 lc -1

#E.g. ------------------example data file -------------------
1.01           1  751    0  1.006    203000 12409  #Fitted_point  monotonic
0.01914      640  421.6  0  0.01707  203000  1280  #Fitted_point manually
0.00968     2300  393.6  0  0.00774  203000   879  #Fitted_point  manually
0.00739     4000  382.1  0  0.00551  203000   757  #Fitted_point  manually
0.00532     8000  367.4  0  0.00351  203000   630  #Fitted_point  manually
0.00424    15000  356.6  0  0.00249  203000   554  #Fitted_point  manually
0.00346    24000  346    0  0.00176  203000   493  #Fitted_point  manually
0.00273    50000  332.1  0  0.0011   203000   429  #Fitted_point  manually
0.00234    80000  321.5  0  0.00076  203000   391  #Fitted_point  manually
0.00204   164000  310.8  0  0.00051  203000   359  #Fitted_point  manually
0.00175   400000  296.1  0  0.00029  203000   324  #Fitted_point  manually
0.00158  1000000  284.3  0  0.00018  203000   302  #Fitted_point  manually
0.00145  2400000  271.9  0  0.00011  203000   283  #Fitted_point  manually
0.00131 15000000  254.6  0  0.00005  203000   260  #Fitted_point  manually
0.00131 20000000  254.6  0  0.00005  203000   260  #Fitted_point  manually


Example plot is attached

gnuplot version is
Version 5.2 patchlevel 2    last modified 2017-11-15
Running opensuse 15.1

best regards,
Al

[Gnuplot-info] Different font size rendering on different machines with pdf terminal

[Helmut H. <he...@ha...>]

Hi,

I use gnuplot 5.2 (patchlevel 8) on various machines running Windows 10 Pro.

When creating plots using the pdf terminal, the font sizes come out 
differently on different machines. Specifically, while
set terminal pdf font "Times-New-Roman,12"
works as expected on one machine, I need to use
set terminal pdf font "Times-New-Roman,6"
on another machine to produce approximately (but not identically) the 
same output on another machine. The plot files are identical otherwise.

Are there some additional settings I'm unaware of that prevent this from 
happening? This is quite annoying because I frequently work on the same 
gnuplot file switching between machines (desktop, laptop) and any help 
in this regard would be greatly appreciated.

Thank you,
--Helmut Haberzettl

[Gnuplot-info] plots combined loop , fucntion, shell command. when comes to epslatex: overlap axis, only one plot in pdf while many in eps file

[Tu C. <tuc...@gm...>]

I have plenty of plots, about 200. using for loop, function, shell
command. 

when I use terminal pdf, all is ok. but I need add some equation on
axis. so I have to use terminal latex or something.

after I changed terminal to epslatex(shown below). 
$ gnuplot **.gplt 
works fine. gives a tex file and eps file. And eps file have several
plots without any text. after I 
$ pdflatex **.tex. 
It only gives a pdf with the first plot and all axis text overlap.

I've trid 
$ set mutliplot
it gives eps and pdf all plots overlapped in one plot. so do any text.

I've searched several days for solution. can't find any demo or
question as complex as I have(loop, function, shell command). also find
none problem like I described above(eps mutliplot while pdf only one).

-----------------------------------------------------------------------
main part of my gnuplot script:

set terminal epslatex size 45cm, 30cm color colortext standalone
set output "result_MVyes_inVno_aver_vg_percent_n-p.tex"

array distance[8] = ["0.1", "0.5", "1.0", "1.5", "2.0", "3.0", "5.0",
"7.0"]
array start[8] = [24,  719,  1414, 2109, 2804, 3499, 4194, 4889]
array end[8]   = [368, 1063, 1758, 2453, 3148, 3843, 4538, 5233]
getData(fName, n) = sprintf("<(sed -n '%d,%dp' %s)", start[n], end[n],
fName)

array clean[3] = ["po", "no", "BGno"]
fileName(year, i) = sprintf("origin_fluid_%s%s.dat", year, clean[i])
comData(n, year) = sprintf("<(paste %s %s |sed -n '%d,%dp')",
fileName(year, 1), fileName(year, 2), start[n], end[n])

uds0 = "using 2:(100*($12-$28)/$28)"

set xlabel 'y'
set key out vert
set logscale x 10
set format y '$%.1e$'
set xrange [10**(-2):3.5]

####################################
# atmospheric pressure             #
#                                  #
####################################

set yrange [-2*10**(-4):2*10**(-4)]
#####################################
set ylabel 'atmospheric pressure'
set title 'every year every distance polluted atmospheric pressure'
plot for [year in "2004 2005 2006"] for [i = 1:8] comData(i, year)
@uds0 with linespoint pn 5 lt (abs(year)-2004)*8+i title year."
".distance[i]

####################################年份不变
set title '2004 every distance polluted and nonpolluted atmospheric
pressure'
plot for [year in "2004"] for [i = 1:8] comData(i, year) @uds0 with
linespoint pn 5 lt (abs(year)-2004)*8+i title year." ".distance[i]

set title '2005 every distance polluted and nonpolluted atmospheric
pressure'
plot for [year in "2005"] for [i = 1:8] comData(i, year) @uds0 with
linespoint pn 5 lt (abs(year)-2004)*8+i title year." ".distance[i]

set title '2006 every distance polluted and nonpolluted atmospheric
pressure'
plot for [year in "2006"] for [i = 1:8] comData(i, year) @uds0 with
linespoint pn 5 lt (abs(year)-2004)*8+i title year." ".distance[i]

####################################
do for [i = 1:8] {
   set title "every year ".distance[i]." polluted and nonpolluted
atmospheric pressure"
   plot for [year in "2004 2005 2006"] comData(i, year) @uds0 with
linespoint pn 5 lt (abs(year)-2003) title year." ".distance[i]
}


-- 
Tu Can <tuc...@gm...>

Re: [Gnuplot-info] appending to default Makefile variable value

[Tait <gnu...@t4...>]

It looks like the .pro file should add: CONFIG += c++11 (at a
minimum, or possibly c++14 or c++17). See
https://wiki.qt.io/How_to_use_C%2B%2B11_in_your_Qt_Projects and
https://stackoverflow.com/questions/19398438/c-qt-how-to-add-std-c11-to-the-makefile-which-is-generated-by-qmake
for example. Can you try that and share whether it fixes the
issue for you?

GCC 4.8.2 is more than seven years old at this point -- roughly
the era of gnuplot 4.6. Any non-ancient version of GCC defaults
to -std=c++14 if unspecified, so this will only affect a select
few people, like yourself, building gnuplot on old systems.


Kevin Klein <kk...@gm...> said (on 2020/09/04):
> Date: Fri, 4 Sep 2020 12:17:07 -0500
> From: Kevin Klein <kk...@gm...>
> To: gnu...@li...
> Subject: [Gnuplot-info] appending to default Makefile variable value
> 
> When running "make" while trying to build gnuplot on RHEL7, I initially
> received the following error:
> 
> In file included from /usr/include/c++/4.8.2/type_traits:35:0,
>                  from /usr/include/qt5/QtCore/qglobal.h:45,
>                  from /usr/include/qt5/QtCore/QtCore:4,
>                  from qterminal/qt_term.cpp:51:
> /usr/include/c++/4.8.2/bits/c++0x_warning.h:32:2: error: #error This file
> requires compiler and library support for the ISO C++ 2011 standard. This
> support is currently experimental, and must be enabled with the -std=c++11
> or -std=gnu++11 compiler options.
> 
> Based on the above message, "-std=c++11" needed to be added to the
> QT_CFLAGS variable that is generated in the Makefile.  Ultimately I did
> this by looking at the default value for this variable in config.log, then
> running the following in the shell:
> 
> setenv QT_CFLAGS '-std=c++11 <QT_CFLAGS>'
> 
> where <QT_CFLAGS> is the default value for QT_CFLAGS as listed in
> config.log, then re-running configure.  This generated a Makefile with
> QT_CFLAGS set to its default value plus " -std=c++11", which then allowed
> the "make" (and "make install") to complete successfully.
> 
> My question is:
> 
> Is there a simple one-liner way to specify that something should be
> appended to the QT_CFLAGS variable when configure is run rather than
> replacing it?  If I had run the following in the shell:
> 
> setenv QT_CFLAGS -std=c++11
> 
> And then run configure, the resulting Makefile would only contain
> "-std=c++11" as the value for QT_CFLAGS rather than that plus all of the
> default library include paths, and "make" would have failed.

[Gnuplot-info] appending to default Makefile variable value

[Kevin K. <kk...@gm...>]

When running "make" while trying to build gnuplot on RHEL7, I initially
received the following error:

In file included from /usr/include/c++/4.8.2/type_traits:35:0,
                 from /usr/include/qt5/QtCore/qglobal.h:45,
                 from /usr/include/qt5/QtCore/QtCore:4,
                 from qterminal/qt_term.cpp:51:
/usr/include/c++/4.8.2/bits/c++0x_warning.h:32:2: error: #error This file
requires compiler and library support for the ISO C++ 2011 standard. This
support is currently experimental, and must be enabled with the -std=c++11
or -std=gnu++11 compiler options.

Based on the above message, "-std=c++11" needed to be added to the
QT_CFLAGS variable that is generated in the Makefile.  Ultimately I did
this by looking at the default value for this variable in config.log, then
running the following in the shell:

setenv QT_CFLAGS '-std=c++11 <QT_CFLAGS>'

where <QT_CFLAGS> is the default value for QT_CFLAGS as listed in
config.log, then re-running configure.  This generated a Makefile with
QT_CFLAGS set to its default value plus " -std=c++11", which then allowed
the "make" (and "make install") to complete successfully.

My question is:

Is there a simple one-liner way to specify that something should be
appended to the QT_CFLAGS variable when configure is run rather than
replacing it?  If I had run the following in the shell:

setenv QT_CFLAGS -std=c++11

And then run configure, the resulting Makefile would only contain
"-std=c++11" as the value for QT_CFLAGS rather than that plus all of the
default library include paths, and "make" would have failed.

[Gnuplot-info] splot zooming crops plot at bounding box of visible axes

[Phil V <pv....@gm...>]

Hello, is there a way to zoom an interactive plot beyond the bounds of
the axes? Or to increase the size of the axes relative to the window?
Currently the image gets cropped off at an invisible bounding box far
away from the edges of the window.


DETAILS:
I am using the command
splot FILE using 1:2:3 with lines

This creates a '3D' wire frame as expected.
I am trying to zoom the image to take up more of the screen.
When I turn the mousewheel it shifts the plot along the y axis.
When I shift turn the mousewheel it shifts the plot along the x axis.
When I control-mousewheel down the axes labels increase in one axis
and the plot scales down visually.
When I control-mousewheel up, the opposite happens, but the plot gets
cropped off at an invisible bounding box.

The magnifying glass buttons allow to reset(1) and the zoom in and out
buttons behave inconsistently: (-) repeated clicks shrinks sometimes
but expands other times. (+) similarly. Sometimes gives the KDE
message "bell in session 'default'".

QUESTIONS:
1. is there a way to zoom an interactive plot beyond the bounds of
the axes?
Q2. is there a way to increase the size of the axes relative to the window?
Ideally cropping would happen only at the edges of the window.

[Gnuplot-info] Plotting several curve using a for loop

[<mar...@ya...>]

Hello everyone, 

I would like to plot several curve using a single line thank to a for loop. 

I currently have : 
plot @chemin_tot_DNS u (@yplusfroid_DNS):(@wT*rho_fr/qwsurcpmf_LES_DNS) @format_dns_fr title "SND", \
@chemin_tot_nomodel u (@yplusfroid_nomodel):(@wT*rho_fr/qwsurcpmf_LES_nomodel) @format_nomodel_fr title "Sans modele (Quick)",  \
@chemin_tot_fsim100sim100 every 1 u (@yplusfroid_fsim100sim100):(@wT*rho_fr/qwsurcpmf_LES_fsim100sim100) pt 1 @pts_fr title "Similarite (Quick)", \
@chemin_tot_fwal055wal055 every 1 u (@yplusfroid_fwal055wal055):(@wT*rho_fr/qwsurcpmf_LES_fwal055wal055) pt 8 @pts_fr title "WALE (Quick)"

plot @chemin_tot_DNS u (@yplusfroid_DNS):(@uT*rho_fr/qwsurcpmf_LES_DNS) @format_dns_fr title "SND", \
@chemin_tot_nomodel u (@yplusfroid_nomodel):(@uT*rho_fr/qwsurcpmf_LES_nomodel) @format_nomodel_fr title "Sans modele (Quick)",  \
@chemin_tot_fsim100sim100 every 1 u (@yplusfroid_fsim100sim100):(@uT*rho_fr/qwsurcpmf_LES_fsim100sim100) pt 1 @pts_fr title "Similarite (Quick)", \
@chemin_tot_fwal055wal055 every 1 u (@yplusfroid_fwal055wal055):(@uT*rho_fr/qwsurcpmf_LES_fwal055wal055) pt 8 @pts_fr title "WALE (Quick)", \


I would like to do something like : 
liste="@chemin_tot_DNS @chemin_tot_nomodel @chemin_tot_fsim100sim100 @chemin_tot_fwal055wal055"
plot for [file in liste] file using (@yplusfroid_VARIABLE):(@wT*rho_fr/qwsurcpmf_LES_VARIABLE)
plot for [file in liste] file using (@yplusfroid_VARIABLE):(@uT*rho_fr/qwsurcpmf_LES_VARIABLE)

where VARIABLE take the value DNS when liste=chemin_tot_DNS, the value nomodel when liste=chemin_tot_nomodel...

I have 2 problem :
1) Gnuplot does not recognize the value @chemin_tot_DNS. This is not so problematic since I can give the value of @chemin_tot_DNS but if you know how to use the @chemin_tot_DNS in the for loop it would be wonderful.
2) I do not succeed in adapting VARIABLE in order that it changes values for each iteration of the for loop. Is there a solution to this issue ? 

Thank you in advance for your answer, 
Sincerely, 
Martin

Re: [Gnuplot-info] rtics NOT starting on 0 degrees

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

glad to hear that this helped a little.
But I again forgot... gnuplot has (this annoying) integer division.

(i*360/16+11.25)   should read  (i*360./16+11.25) or (i*360/16.+11.25)
Otherwise i*360/16 would be an integer.

Re: [Gnuplot-info] rtics NOT starting on 0 degrees

[Kevin O. <rko...@gm...>]

On Sun, Jul 12, 2020 at 1:40 PM theozh <th...@gm...> wrote:

> I'm not sure whether you can do this directly. I assume you mean the ttics.
> Although you can set a polar grid angle (check 'help grid') you can only
> start at multiples of 90 degrees. Check 'help theta'.
> A workaround would be to draw the gridlines manually.
> I hope this still helps somehow.
> Best, Theo.
>
> Try the following example:
>
> ### polar plot with grids
> reset session
> set size square
> set angle degrees
>
> set polar
> set theta clockwise top
> unset border
> unset tics
> unset raxis
> unset key
> set border polar
> set ttics 11.25, 22.50 format ""
> set rtics 10 scale 0 format ""
> set grid rtics
>
> set rrange[0:40]
> # manually draw grid lines
> do for [i=1:16] {
>     set arrow i from 0,0 to polar (i*360/16+11.25),40 lw 0.5 dt 3 lc rgb
> "black" nohead
> }
>
> set samples 16
> plot t/9 w lp pt 7 lc rgb "red"
> ### end of code
>
Thank you, Theo. I have not tried drawing the grid, but your suggestions
have gotten me very close to what I'm looking for. Especially for 'set
theta" and the use of "format" on set ttics". Since the manual for ttics
explicitly stated "The angular position is always labeled in degrees.", I
didn't even try format. As a result, I now am close to my goal. Still a
couple of things I need to tweak... like playing with the smoothing.

Again, thanks! If you are curious, it's at
http://ykoberman.mooo.com/display-weather-graph.pl?Plot=dir
Not too interesting unless you live near Lawrence Livermore National Lab,
but it's all gnuplot except the main page (Home).
--
Kevin Oberman, Part time kid herder and retired Network Engineer
E-mail: rko...@gm...
PGP Fingerprint: D03FB98AFA78E3B78C1694B318AB39EF1B055683

Re: [Gnuplot-info] rtics NOT starting on 0 degrees

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

I'm not sure whether you can do this directly. I assume you mean the ttics.
Although you can set a polar grid angle (check 'help grid') you can only start at multiples of 90 degrees. Check 'help theta'.
A workaround would be to draw the gridlines manually.
I hope this still helps somehow.
Best, Theo.

Try the following example:

### polar plot with grids
reset session
set size square
set angle degrees

set polar
set theta clockwise top
unset border
unset tics
unset raxis
unset key
set border polar
set ttics 11.25, 22.50 format ""
set rtics 10 scale 0 format ""
set grid rtics

set rrange[0:40]
# manually draw grid lines
do for [i=1:16] {
    set arrow i from 0,0 to polar (i*360/16+11.25),40 lw 0.5 dt 3 lc rgb "black" nohead
}

set samples 16
plot t/9 w lp pt 7 lc rgb "red"
### end of code

[Gnuplot-info] rtics NOT starting on 0 degrees

[Kevin O. <rko...@gm...>]

I am trying to plot a compass type polar plot. As such, I want grid lines
at 22.5 degree intervals, but I want them to start 11.25 degrees before or
after 0 degrees. I've been studying the manual, but have been unable to
figure it out.

To fully clarify, I want grid lines between N, NNE, NE, ENE, E, etc. with
one at 11.25 degrees before and after each section.

Is there a clever way to do this? Or, maybe an obvious one I missed?
--
Kevin Oberman, Part time kid herder and retired Network Engineer
E-mail: rko...@gm...
PGP Fingerprint: D03FB98AFA78E3B78C1694B318AB39EF1B055683

Re: [Gnuplot-info] How to plot multiple fit functions in a single box

[rekha s. <rek...@gm...>]

Dear Sir,
I corrected that mistake but still I am getting error which is pointing
towards:
% g +  a_1/T

Attached are test.data and test.gnu file (part of test.gnu is mentioned
below).
f(x,a0,a1) = a0 + a1/x
fit f(x,a01,a11) 'test.data' using 1:2 via a01,a11
fit f(x,a02,a12) 'test.data' using 1:3 via a02,a12
fit f(x,a03,a13) 'test.data' using 1:4 via a03,a13
fit f(x,a04,a14) 'test.data' using 1:5 via a04,a14
fit f(x,a05,a15) 'test.data' using 1:6 via a05,a15
fit f(x,a06,a16) 'test.data' using 1:7 via a06,a16
fit f(x,a07,a17) 'test.data' using 1:8 via a07,a17
fit f(x,a08,a18) 'test.data' using 1:9 via a08,a18
plot \
   'test.data' using 1:2 w points pt 1 t  , \
    f(x,a01,a11) t sprintf("K_{fit} = % g +  a_1/T", a01), \
   '' using 1:3 w points pt 1 t  , \
    f(x,a02,a12) t sprintf("K_{fit} = % g +  a_1/T", a02), \
   '' using 1:4 w points pt 1 t  , \
    f(x,a03,a13) t sprintf("K_{fit} = % g +  a_1/T", a03), \
  '' using 1:5 w points pt 1 t  , \
    f(x,a04,a14) t sprintf("K_{fit} = % g +  a_1/T", a04), \
  '' using 1:6 w points pt 1 t  , \
    f(x,a05,a15) t sprintf("K_{fit} = % g +  a_1/T", a05), \
  '' using 1:7 w points pt 1 t  , \
    f(x,a06,a16) t sprintf("K_{fit} = % g +  a_1/T", a06), \
  '' using 1:8 w points pt 1 t  , \
    f(x,a07,a17) t sprintf("K_{fit} = % g +  a_1/T", a07), \
  '' using 1:9 w points pt 1 t  , \
    f(x,a08,a18) t sprintf("K_{fit} = % g +  a_1/T", a08), \

On Wed, Jul 8, 2020 at 1:26 AM Hans-Bernhard Bröker <HBB...@t-...>
wrote:

> Am 04.07.2020 um 21:57 schrieb rekha sharma:
> > Dear Sir
> > My gnuplot version is
> >   Version 5.2 patchlevel 2    last modified 2017-11-01
>
> That's getting a bit stale, bit it should still be usable.
>
> >
> > With the  test.gnu file as suggested by you
>
> ... I'm afraid you did miss a few important aspects of the proposal...
>
> > this  ^ points to a08) .
>
> Now look at the version you put together and ask yourself: is a08 really
> treated the same as, say, a03, and a18 or is there possibly some
> difference that might explain why the error concerned a08, but not any
> of the other 15 variables?
>


-- 




Best wishes

Re: [Gnuplot-info] How to plot multiple fit functions in a single box

[Hans-Bernhard B. <HBB...@t-...>]

Am 04.07.2020 um 21:57 schrieb rekha sharma:
> Dear Sir
> My gnuplot version is
>   Version 5.2 patchlevel 2    last modified 2017-11-01

That's getting a bit stale, bit it should still be usable.

> 
> With the  test.gnu file as suggested by you

... I'm afraid you did miss a few important aspects of the proposal...

> this  ^ points to a08) .

Now look at the version you put together and ask yourself: is a08 really 
treated the same as, say, a03, and a18 or is there possibly some 
difference that might explain why the error concerned a08, but not any 
of the other 15 variables?

Re: [Gnuplot-info] How to plot multiple fit functions in a single box

[rekha s. <rek...@gm...>]

Dear Sir
My gnuplot version is
 Version 5.2 patchlevel 2    last modified 2017-11-01

With the  test.gnu file as suggested by you
f(x,a0,a1) = a0 + a1/x

fit f(x,a01,a11) 'test.data' using 1:2 via a01,a11
fit f(x,a02,a12) 'test.data' using 1:3 via a02,a12
fit f(x,a02,a13) 'test.data' using 1:4 via a02,a13
fit f(x,a02,a14) 'test.data' using 1:5 via a02,a14
fit f(x,a02,a15) 'test.data' using 1:6 via a02,a15
fit f(x,a02,a16) 'test.data' using 1:7 via a02,a16
fit f(x,a02,a17) 'test.data' using 1:8 via a02,a17
fit f(x,a02,a18) 'test.data' using 1:9 via a02,a18
plot \
   'test.data' using 1:2 w points pt 1 t  , \
    f(x,a01,a11) t sprintf("K_{fit} = % g +  a_1/T", a01), \
   '' using 1:3 w points pt 1 t  , \
    f(x,a02,a12) t sprintf("K_{fit} = % g +  a_1/T", a02), \
   '' using 1:4 w points pt 1 t  , \
    f(x,a03,a13) t sprintf("K_{fit} = % g +  a_1/T", a03), \
  '' using 1:5 w points pt 1 t  , \
    f(x,a04,a14) t sprintf("K_{fit} = % g +  a_1/T", a04), \
  '' using 1:6 w points pt 1 t  , \
    f(x,a05,a15) t sprintf("K_{fit} = % g +  a_1/T", a05), \
  '' using 1:7 w points pt 1 t  , \
    f(x,a06,a16) t sprintf("K_{fit} = % g +  a_1/T", a06), \
  '' using 1:8 w points pt 1 t  , \
    f(x,a07,a17) t sprintf("K_{fit} = % g +  a_1/T", a07), \
  '' using 1:9 w points pt 1 t  , \
  '' using 1:4 w points pt 1 t  , \
    f(x,a03,a18) t sprintf("K_{fit} = % g +  a_1/T", a08), \

I am getting error below error
                                               ^
"hans.gnu", line 27: invalid expression

this  ^ points to a08) .

Regards
Rekha


On Sun, Jul 5, 2020 at 12:58 AM Hans-Bernhard Bröker <HBB...@t-...>
wrote:

> Am 04.07.2020 um 17:04 schrieb rekha sharma:
>
> > f(x) = (a0 + a1/x)
> > fit f(x) 'test.data' using 1:2 via a0,a1
> > plot 'test.data' using 1:2 w points pt 1 t  ,  f(x) t sprintf("K_{fit} =
> > a_0 +  a_1/T", a0)
> [... and the same for columns 3 to 9 as the y axis]
>
> > Using this code, I am getting a plot in eight boxes
>
> Using only that code, that cannot happen.  What you would get is 8
> plots, but not in 8 boxes, but rather on 8 separate pages of output (if
> your output supports that), or just a single plot window being updated 8
> times.  To get "a plot in 8 boxes", you would have to have "set
> multiplot layout ..." in effect somehow.
>
> > Is there any workaround for this issue in gnuplot so that I can plot all
> > above in a single box?
>
> For all the above to show up in a single "box", i.e. on a single page,
> you have to change the the way you define your fit model function and
> its parameters.  And then you have to undo the parts of your scripts
> that you claim aren't there, about "set multiplot"
>
> f(x,a0,a1) = a0 + a1/x
>
> fit f(x,a01,a11) 'test.data' using 1:2 via a01,a11
> fit f(x,a02,a11) 'test.data' using 1:3 via a02,a12
> [...]
>
> plot \
>    'test.data' using 1:2 w points pt 1 t  , \
>     f(x,a01,a11) t sprintf("K_{fit} = % g +  a_1/T", a01), \
>    '' using 1:3 w points pt 1 t  , \
>     f(x,a02,a12) t sprintf("K_{fit} = % g +  a_1/T", a02), \
>    '' using 1:4 w points pt 1 t  , \
>     f(x,a03,a13) t sprintf("K_{fit} = % g +  a_1/T", a03), \
> [...]
>
> If your gnuplot is new enough, the above can be improved upon using for
> loops for the fits and the parts of the plot, and arrays for the
> parameters.  And you'll want to re-think that sprintf, I think.
>
> That will give you a rather packed diagram, though.
>


-- 




Best wishes

Ms. Rekha
Teaching Assistant,
Department of Physics
RK college, Jaipur, India
Mob.: +11 90-95 790 71 697
Email: rek...@gm...

Re: [Gnuplot-info] How to plot multiple fit functions in a single box

[Hans-Bernhard B. <HBB...@t-...>]

Am 04.07.2020 um 17:04 schrieb rekha sharma:

> f(x) = (a0 + a1/x)
> fit f(x) 'test.data' using 1:2 via a0,a1
> plot 'test.data' using 1:2 w points pt 1 t  ,  f(x) t sprintf("K_{fit} =
> a_0 +  a_1/T", a0)
[... and the same for columns 3 to 9 as the y axis]

> Using this code, I am getting a plot in eight boxes 

Using only that code, that cannot happen.  What you would get is 8 
plots, but not in 8 boxes, but rather on 8 separate pages of output (if 
your output supports that), or just a single plot window being updated 8 
times.  To get "a plot in 8 boxes", you would have to have "set 
multiplot layout ..." in effect somehow.

> Is there any workaround for this issue in gnuplot so that I can plot all
> above in a single box?

For all the above to show up in a single "box", i.e. on a single page, 
you have to change the the way you define your fit model function and 
its parameters.  And then you have to undo the parts of your scripts 
that you claim aren't there, about "set multiplot"

f(x,a0,a1) = a0 + a1/x

fit f(x,a01,a11) 'test.data' using 1:2 via a01,a11
fit f(x,a02,a11) 'test.data' using 1:3 via a02,a12
[...]

plot \
   'test.data' using 1:2 w points pt 1 t  , \
    f(x,a01,a11) t sprintf("K_{fit} = % g +  a_1/T", a01), \
   '' using 1:3 w points pt 1 t  , \
    f(x,a02,a12) t sprintf("K_{fit} = % g +  a_1/T", a02), \
   '' using 1:4 w points pt 1 t  , \
    f(x,a03,a13) t sprintf("K_{fit} = % g +  a_1/T", a03), \
[...]

If your gnuplot is new enough, the above can be improved upon using for 
loops for the fits and the parts of the plot, and arrays for the 
parameters.  And you'll want to re-think that sprintf, I think.

That will give you a rather packed diagram, though.

Re: [Gnuplot-info] How to plot multiple fit functions in a single box

[rekha s. <rek...@gm...>]

Thank you theozh Sir,
I am lucky to get your contact information.
You have helped me a lot for most of my stuff.

Dear Norwid Sir,
I have already tried your trick f(x), g(x), etc. but it could not help me.
I have
I have tried to put all the plots and fits into a single plot command (an
example in the test.gnu file is mentioned) and it gave me error
"test.gnu", line 23: undefined variable: fit

Here are my data files

https://we.tl/t-kOdOPfzrcK

<https://we.tl/t-kOdOPfzrcK>
I need all data in a single block. For a sample plot (it is for a single
curve) I have kept a standared-plot.png file in the above data.

regards
Rekha

On Sat, Jul 4, 2020 at 9:59 PM theozh <th...@gm...> wrote:

> check the answer there:
>
>
> https://stackoverflow.com/questions/62726501/how-to-merge-all-plots-in-one-window-in-gnuplot
>
>
>
> _______________________________________________
> gnuplot-info mailing list
> gnu...@li...
> Membership management via:
> https://lists.sourceforge.net/lists/listinfo/gnuplot-info
>


-- 




Best wishes

Ms. Rekha
Teaching Assistant,
Department of Physics
RK college, Jaipur, India
Mob.: +11 90-95 790 71 697
Email: rek...@gm...

Re: [Gnuplot-info] How to plot multiple fit functions in a single box

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

check the answer there:

https://stackoverflow.com/questions/62726501/how-to-merge-all-plots-in-one-window-in-gnuplot

Re: [Gnuplot-info] How to plot multiple fit functions in a single box

[Norwid B. <nb...@ya...>]

Dear Rekha,

my understanding is that you have one file of data and the script may
perform the fits sequentially (using 1:2, using 1:3, using 1:4, etc.).
The structure of the underlying function seems to be the same for all of
them, too, f(x) = (a0 + a1/x).

May renaming the functions be an option for you, e.g. f1(x), f2(x), f3(x),
thus creating to discernible fits?  I speculate this requires some change
in the script used (the example covers only two data):

f1(x) = (a0 + a1/x)
fit f1(x) 'test.data' using 1:2 via a0,a1
plot 'test.data' using 1:2 w points pt 1 t  ,  f1(x) t sprintf("K_{fit} =
a_0 +  a_1/T", a0)

f2(x) = (a0 + a1/x)
fit f2(x) 'test.data' using 1:3 via a0,a1
plot  'test.data' using 1:3 w points pt 1 t   ,  f2(x) t sprintf("K_{fit} =
a_0 +  a_1/T", a0)

Using the integer count may yield a script easier to maintain and expand
than using a set of functions named with changing letters, e.g. f(x), g(x),
h(x), etc.

If so, because of similarity in these blocks and the predictable changes
between them, maybe this may be parameterized once; you would have one
block of definition only.  When calling for this function, its sole variable
then were n to perform n fits / plot n curves; there wouldn't be a need
to write n blocks (dry versus wet programming principle).

Norwid