gnuplot-info Mailing List for gnuplot (Page 15)

In section "activities", the sourceforge page[1] lists
gnuplot/5.4.3/gnuplot-5.4.3.tar.gz as an update, just 5h ago. 

Norwid

[1] https://sourceforge.net/projects/gnuplot/

Today on cygwin64 setup, I found 5.4.3 binaries were prepared.

Then I accessed to the SourceForge page.
https://sourceforge.net/projects/gnuplot/files/gnuplot/5.4.3/

But I could not access any files in the page.

What happens?

Tatsuro

On 2021-11-23 22:30, Dave Horsfall wrote:

> I now have a working inkjet (it took a while to configure it for my 
> LAN)
> and it occurred to me that if I choose the RGB colours appropriately 
> for
> each "time slot" then not only will they be intuitive on an inkjet (I
> envisage a sort of a light colour getting brighter then fading to dark 
> in
> the form of coloured bars) but will also show up as an appropriate
> greyscale on the laser printer so that I can use either one.

[I'm new to gnuplot, and haven't actually used it yet, but have looked 
at
lots of examples of its use.  I'm expecting to write programs to take
data from elsewhere and generate gnuplot scripts to create charts, also
health-related.  So I've got no practical advice yet.]


Even so... reading this I wondered how much data you were planning to 
put
on each graph?  Just one day's worth, or many days'?

Is the colour gradient supposed to represent time of day, or eg the
width of a (say) 2-hour band of time in the day - both of which are what
I thought you might mean (but don't see the point), or to show levels of
danger from seriously hypo, mildly hypo, normal, hyper etc?

It seems to me that colour, if it's worth using at all, is worth using
to highlight the verging on dangerous or actually dangerous readings.


Regarding your other question about times of day, what's wrong (though 
it
maybe depends on how many days' data are in one chart) with just using
the hour number (from 00 to 23)?

-- 
Jeremy Nicoll - my opinions are my own

I made a directory bin/platforms and copy there qwindows.dll.
gnuplot with qt starred without problem.

Tatsuro

> ----- Original Message -----

> 
> I tried to build gnuplot from souce on windows.
> Build itself was successful and windows, wxt, and qt terminals works if gnuplot is executed in config/mingw directory.
> However, if binary files (program and dynamic link library files) are installed, I was not able to start qt terminal due to the error below
> 
> This application failed to start because no Qt platform plugin
> could be initialized. Reinstalling the application may fix this
> problem.
> 
> Screenshot
> http://tmacchant33.starfree.jp/Files/gnuplot_qt_init.png
> 
> It will be grateful for me if answers or suggestions will be sent.
> 
> Tatsuro
> 
> 
> 
> _______________________________________________
> gnuplot-info mailing list
> gnu...@li...
> Membership management via: https://lists.sourceforge.net/lists/listinfo/gnuplot-info
> 

I tried to build gnuplot from souce on windows.
Build itself was successful and windows, wxt, and qt terminals works if gnuplot is executed in config/mingw directory.
However, if binary files (program and dynamic link library files) are installed, I was not able to start qt terminal due to the error below

This application failed to start because no Qt platform plugin
could be initialized. Reinstalling the application may fix this
problem.

Screenshot
http://tmacchant33.starfree.jp/Files/gnuplot_qt_init.png

It will be grateful for me if answers or suggestions will be sent.

Tatsuro

I would like to add '-5.5' to gnuplot and gnuplot-x11 (i.e. gnuplot-5.5 and gnuplot-x11-5.5)
./.configure --help

Program names:
  --program-prefix=PREFIX            prepend PREFIX to installed program names
  --program-suffix=SUFFIX            append SUFFIX to installed program names
  --program-transform-name=PROGRAM   run sed PROGRAM on installed program names

I tried --program-suffix=SUFFIX but results are not expected.

Please show me thw way.

Tatsuro

On Mon, 22 Nov 2021, Hernán De Angelis wrote:

> Gnuplot can handle many different formats for time data and plot. Have a 
> look at pages 58 and 151-152 in the pdf manual. May be you find one that 
> suits you.

Ick; I don't really want to dig through that right now (I'll save it for a 
rainy day), but I did have another idea; please note that I am neither 
asking for nor providing medical advice...

I now have a working inkjet (it took a while to configure it for my LAN) 
and it occurred to me that if I choose the RGB colours appropriately for 
each "time slot" then not only will they be intuitive on an inkjet (I 
envisage a sort of a light colour getting brighter then fading to dark in 
the form of coloured bars) but will also show up as an appropriate 
greyscale on the laser printer so that I can use either one.

Now, graphical arts is not exactly my forté, so what would be suitable RGB 
codes?  At worst I can always dig out a colour wheel and muck around a 
bit...

In other words, I am trying to map RGB colours into corresponding shades 
of grey if that makes sense.

Thanks.

-- Dave

Gnuplot can handle many different formats for time data and plot. Have a 
look at pages 58 and 151-152 in the pdf manual. May be you find one that 
suits you.

/H.


Den 2021-11-21 kl. 20:18, skrev Dave Horsfall:
> I was recently diagnosed with diabetes (ugh!) and not only do I need 
> to track my glucose levels but also the time of day; the latter is 
> just as important, involving a test two hours after a meal.  The times 
> of the meals are also important; it's not your "Western" diet i.e. 
> three meals/day, but what I may as well call the "Eastern" diet i.e. 
> snacking throughout the day with no major meal.
>
> My GP wants to see a table of this, but I lean towards a plot instead; 
> the application that came with my meter is a bit tricky to interpret.
>
> The problem is how to represent the time; I can think of three ways:
>
> 1) A fractional day; hard to compute and interpret.
>
> 2) Coloured bars; I don't have a working inkjet.
>
> 3) A 3-D graph with the Z axis as the time, resulting in a wiggly line.
>
> Any ideas on how to do it?
>
> Thanks.
>
> -- Dave
>
>
> _______________________________________________
> gnuplot-info mailing list
> gnu...@li...
> Membership management via: 
> https://lists.sourceforge.net/lists/listinfo/gnuplot-info

I was recently diagnosed with diabetes (ugh!) and not only do I need to 
track my glucose levels but also the time of day; the latter is just as 
important, involving a test two hours after a meal.  The times of the 
meals are also important; it's not your "Western" diet i.e. three 
meals/day, but what I may as well call the "Eastern" diet i.e. snacking 
throughout the day with no major meal.

My GP wants to see a table of this, but I lean towards a plot instead; the 
application that came with my meter is a bit tricky to interpret.

The problem is how to represent the time; I can think of three ways:

1) A fractional day; hard to compute and interpret.

2) Coloured bars; I don't have a working inkjet.

3) A 3-D graph with the Z axis as the time, resulting in a wiggly line.

Any ideas on how to do it?

Thanks.

-- Dave

Hello,

Here a set of data (file.asc)

0 34.42904
1 34.42924
2 34.42965
3 34.43026
4 34.43106
5 34.43207
6 34.43329
7 34.43469
8 34.43632
9 34.43814
10 34.44017

I wish to fit them with
dep (x, a) = x * a

I can do
1) 
fit Off + dep (x, slope) "file.asc" u ($1*($1+1)):2 via Off, slope
and I get
final sum of squares of residuals : 2.94071e-10
Off            = 34.429           +/- 2.581e-06    (7.496e-06%)
slope           = 0.000101126      +/- 4.802e-08    (0.04749%)

with
Off = 34.429039486

2)
set table $DATA; plot "file.asc" u ($1*($1+1)):($2) "%lf %lf" with table ; unset table
fit Off + dep (x, slope) $DATA u 1:2 via Off, slope
and I get
final sum of squares of residuals : 1.36307e-08
Off            = 34.429           +/- 1.757e-05    (5.103e-05%)
slope           = 0.000101045      +/- 3.27e-07     (0.3236%)

with
Off = 34.4290491247


Option 1) is correct, but not option 2)

Why?

Thanks

===========================================================================
 Patrick DUPRÉ                                 | | email: pd...@gm...
 Laboratoire interdisciplinaire Carnot de Bourgogne
 9 Avenue Alain Savary, BP 47870, 21078 DIJON Cedex FRANCE
 Tel: +33 (0)380395988                    | | Room# D114A
===========================================================================

Hello,

If I do

set table "out_HW.asc"; plot [-10.5:10.5] '+' u 1:(HW (x, mu_0, C, D, E)); unset table
I get a header, like:
# Curve 0 of 1, 1051 points
# Curve title: "'+' u 1:(HW (x, mu_0, C, D, E))"
# x y type
-10.5  0.0667414  i
.....

If I do:
set table "out_HW.asc"; plot [-10.5:10.5] '+' u 1:(HW (x, mu_0, C, D, E)) with table ; unset table 

I get:
 -10.5	 0.0667414
 -10.48	 0.0667415
.....

How can I get also the header with this format?

===========================================================================
 Patrick DUPRÉ                                 | | email: pd...@gm...
 Laboratoire interdisciplinaire Carnot de Bourgogne
 9 Avenue Alain Savary, BP 47870, 21078 DIJON Cedex FRANCE
 Tel: +33 (0)380395988                    | | Room# D114A
===========================================================================

On 18/10/2021 09:13, Romain Gontero wrote:
> Hello,
>
> I'm a PhD student in modeling mechanical behavior of
> materials. For my work, i need to fit a surface on a 3d curve and i do
> it with gnuplot which is awesome. But now i need to do it by myself, i
> need to create in Fortran, an algorithm which do the same thing as
> gnuplot and i would like to know if it is possible to have a look at the
> code doing the fit procedure in gnuplot, and how to do that ?
>
> Thank
> you for your help.
>
> Best regards.
>
> Romain

I suggest you do not try to 'reinvent the wheel'.

Could you send commands to gnuplot using a pipe? That way. you can 
generate your plots using essentially a bunch of Fortran WRITE statements.

Peter

>
>    
> _______________________________________________
> gnuplot-info mailing list
> gnu...@li...
> Membership management via: https://lists.sourceforge.net/lists/listinfo/gnuplot-info

Romain,

if you want to address the computational part of the problem in
Fortran, and subsequently use gnuplot «just» to visualize the
difference plot between the recorded surface and the one fitted,
consider to address your question on the lively public discussion
board of fortran-lang.org, which is

https://fortran-lang.discourse.group/

Norwid

Hello, 

I'm a PhD student in modeling mechanical behavior of
materials. For my work, i need to fit a surface on a 3d curve and i do
it with gnuplot which is awesome. But now i need to do it by myself, i
need to create in Fortran, an algorithm which do the same thing as
gnuplot and i would like to know if it is possible to have a look at the
code doing the fit procedure in gnuplot, and how to do that ? 

Thank
you for your help. 

Best regards. 

Romain 

  

On Friday, 8 October 2021 03:17:02 PDT theozh wrote:
> Dear Ruda,
> 
> yes, this is my current understanding.
> Actually, the improved solution Friedrich proposed (https://stackoverflow.com/a/62170512/7295599) based on my initial workaround is (currently) the most convenient option I know of. As I understand, you want to zoom in/out interactively into indiviual sub-plots.
> Maybe one could think of a very ugly and lengthy workaround, depending on the layout of your multiplot.
> But in general, I don't see this option without changes in the gnuplot source code. For this you have to ask the developers of gnuplot.

The basic limitation that hits "set multiplot" is that
gnuplot does not keep a history of session state;
it only stores the state of the current plot, i.e. the
result of the most recent 'plot' or 'splot' command.

This limitation can be bypassed for web-based (actually
browser-based) plot display applications, because you can
embed more than one plot in a single xml/html page without
using "set multiplot".  The gnuplot demo pages for
svg and canvas terminal are an example of this.

   http://gnuplot.sourceforge.net/demo_svg_5.5/
   http://gnuplot.sourceforge.net/demo_canvas_5.5/

In these cases the zooming is handled by a separate snippet of
javascript code for each plot including the current zoom state.
Gnuplot itself doesn't know anything about it obviously, since the
plots are displayed by a separate program.  The demo pages are
not laid out in a grid like typical multiplot output but
that's just a choice made by the script that creates them;
you could arrange the plots in a grid if you prefer.
The canvas terminal output support could be supplemented by
hot-key handling. The demo pages do not show an example of that.
I have not figured out how to present hot keys for svg output;
suggestions there would be welcome.

One more thought:

Although gnuplot does not keep a history of program state,
is does keep a history of commands.  You can reproduce a
full multiplot rather laboriously by stepping back through the
command history and re-executing the commands one-by-one that
originally created it.  A possible new program feature would
be to internally flag the first and last commands of a
multiplot sequence and allow the user to re-execute the
comple set in one go.  That would at least provide the
equivalent of a "remultiplot" command.
Unfortunately it wouldn't help with zooming.

	Ethan

> best,
> Theo.
> 
> Am 08.10.2021 um 09:25 schrieb Rudolf Sykora:
> > Dear Theozh,
> >
> > theozh <th...@gm...> writes:
> >
> >> Maybe this is helpful to you...
> >> https://stackoverflow.com/a/62166263/7295599
> >
> > thanks for your reply. So the answer is mostly that gnuplot is not
> > smoothly capable of what I want, and what you show in your post is kind
> > of a workaround (which I do not like much).
> >
> > Thank you.
> > Best regards
> > Ruda
> >
> >>
> >> Am 07.10.2021 um 12:04 schrieb Rudolf Sykora via gnuplot-info:
> >>> Hello,
> >>>
> >>>
> >>> I could not find an answer anywhere...
> >>>
> >>> I use the x11 terminal with a multiplot to draw several graphs. Is it
> >>> possible to zoom in/out the individual graphs (with mouse 3rd button,
> >>> keypresses of u/a/n/p etc.) At this moment, when I try to zoom in one
> >>> graph, the zooming is performed but the graph suddenly occupies the
> >>> whole canvas and I cannot see the other graphs of the multiplot.
> >>>
> >>>
> >>> Thanks for comments!
> >>>
> >>> Best regards,
> >>> Rudolf Sykora
> >>
> >>
> >>
> >> _______________________________________________
> >> gnuplot-info mailing list
> >> gnu...@li...
> >> Membership management via: https://lists.sourceforge.net/lists/listinfo/gnuplot-info
> >
> 
> 
> _______________________________________________
> gnuplot-info mailing list
> gnu...@li...
> Membership management via: https://lists.sourceforge.net/lists/listinfo/gnuplot-info
> 

Dear Ruda,

yes, this is my current understanding.
Actually, the improved solution Friedrich proposed (https://stackoverflow.com/a/62170512/7295599) based on my initial workaround is (currently) the most convenient option I know of. As I understand, you want to zoom in/out interactively into indiviual sub-plots.
Maybe one could think of a very ugly and lengthy workaround, depending on the layout of your multiplot.
But in general, I don't see this option without changes in the gnuplot source code. For this you have to ask the developers of gnuplot.

best,
Theo.

Am 08.10.2021 um 09:25 schrieb Rudolf Sykora:
> Dear Theozh,
>
> theozh <th...@gm...> writes:
>
>> Maybe this is helpful to you...
>> https://stackoverflow.com/a/62166263/7295599
>
> thanks for your reply. So the answer is mostly that gnuplot is not
> smoothly capable of what I want, and what you show in your post is kind
> of a workaround (which I do not like much).
>
> Thank you.
> Best regards
> Ruda
>
>>
>> Am 07.10.2021 um 12:04 schrieb Rudolf Sykora via gnuplot-info:
>>> Hello,
>>>
>>>
>>> I could not find an answer anywhere...
>>>
>>> I use the x11 terminal with a multiplot to draw several graphs. Is it
>>> possible to zoom in/out the individual graphs (with mouse 3rd button,
>>> keypresses of u/a/n/p etc.) At this moment, when I try to zoom in one
>>> graph, the zooming is performed but the graph suddenly occupies the
>>> whole canvas and I cannot see the other graphs of the multiplot.
>>>
>>>
>>> Thanks for comments!
>>>
>>> Best regards,
>>> Rudolf Sykora
>>
>>
>>
>> _______________________________________________
>> gnuplot-info mailing list
>> gnu...@li...
>> Membership management via: https://lists.sourceforge.net/lists/listinfo/gnuplot-info
>

Dear Theozh,

theozh <th...@gm...> writes:

> Maybe this is helpful to you...
> https://stackoverflow.com/a/62166263/7295599

thanks for your reply. So the answer is mostly that gnuplot is not
smoothly capable of what I want, and what you show in your post is kind
of a workaround (which I do not like much).

Thank you.
Best regards
Ruda

>
> Am 07.10.2021 um 12:04 schrieb Rudolf Sykora via gnuplot-info:
>> Hello,
>>
>>
>> I could not find an answer anywhere...
>>
>> I use the x11 terminal with a multiplot to draw several graphs. Is it
>> possible to zoom in/out the individual graphs (with mouse 3rd button,
>> keypresses of u/a/n/p etc.) At this moment, when I try to zoom in one
>> graph, the zooming is performed but the graph suddenly occupies the
>> whole canvas and I cannot see the other graphs of the multiplot.
>>
>>
>> Thanks for comments!
>>
>> Best regards,
>> Rudolf Sykora
>
>
>
> _______________________________________________
> gnuplot-info mailing list
> gnu...@li...
> Membership management via: https://lists.sourceforge.net/lists/listinfo/gnuplot-info

Hello,

If I do
plot [0.1:] f(x)
I get scale from 0.1 to the maximum value of the plot
but values lower than 0.1 are possibly plotted.
Is there a way to really trim the values at 0.1?

Thanks

===========================================================================
 Patrick DUPRÉ                                 | | email: pd...@gm...
 Laboratoire interdisciplinaire Carnot de Bourgogne
 9 Avenue Alain Savary, BP 47870, 21078 DIJON Cedex FRANCE
 Tel: +33 (0)380395988                    | | Room# D114A
===========================================================================

Maybe this is helpful to you...
https://stackoverflow.com/a/62166263/7295599

Am 07.10.2021 um 12:04 schrieb Rudolf Sykora via gnuplot-info:
> Hello,
>
>
> I could not find an answer anywhere...
>
> I use the x11 terminal with a multiplot to draw several graphs. Is it
> possible to zoom in/out the individual graphs (with mouse 3rd button,
> keypresses of u/a/n/p etc.) At this moment, when I try to zoom in one
> graph, the zooming is performed but the graph suddenly occupies the
> whole canvas and I cannot see the other graphs of the multiplot.
>
>
> Thanks for comments!
>
> Best regards,
> Rudolf Sykora

Hello,


I could not find an answer anywhere...

I use the x11 terminal with a multiplot to draw several graphs. Is it
possible to zoom in/out the individual graphs (with mouse 3rd button,
keypresses of u/a/n/p etc.) At this moment, when I try to zoom in one
graph, the zooming is performed but the graph suddenly occupies the
whole canvas and I cannot see the other graphs of the multiplot.


Thanks for comments!

Best regards,
Rudolf Sykora

On Thu, 16 Sep 2021, Patrick Dupre wrote:

> As a result of the fit, I get:
> Offset          = -2.48586e+07
> How can I display more digits?

The number of digits you get in that message is being controlled by a
hard-coded format string on line 1171 or line 1200 of fit.c in the
(commit a5562b) Gnuplot source code.  But one thing you could do is,
straight after the fit is completed, issue the command "print Offset",
which will provide a greater number of digits (controlled by a
different hard-coded format string, on line 1576 of util.c).

_However_: think carefully about whether printing any more digits is
useful, given that there will be some standard error in the inferred
value of the parameter.

On Mon, 20 Sep 2021, Gabriel Dupras wrote:

[...]

> Gnuplot makes a distinction between missing data and invalid data (e.g. 
> "NaN", 1/0.).  For example invalid data causes a gap in a line drawn 
> through sequential data points; missing data does not.

I'd forgotten that; thanks.  On reflection, interpolating missing data is 
indeed the right thing to do (and to think that I gained a Distinction in 
statistics way back in my university days).

> So you should *not* set "x" as the string for missing data. Instead, try 
> this:

[...]

Perfect; that's exactly what I want!  Thanks; I'm not as familiar with 
Gnuplot as I used to be (I really must buy the book).

-- Dave

On Mon, Sep 20, 2021 at 3:45 PM Dave Horsfall <da...@ho...> wrote:

> Anyone?
>
> -- Dave
>
> ---------- Forwarded message ----------
> Date: Tue, 31 Aug 2021 08:37:08 +1000 (EST)
> From: Dave Horsfall <da...@ho...>
> To: Gnuplot Mailing List <gnu...@li...>
> Subject: Gnuplot is interpolating between missing data
>
> Gnuplot (on a Mac), 5.4p2
>
> I have some missing data which I indicate with "x" (set datafile missing
> "x"),
> yet Gnuplot seems to be drawing a straight line between the valid data.
> Shouldn't it be blank instead?
>
> -- Dave
>
Not having seen an answer, I'll give it a shot. I believe that this is the
default for gnuplot and that having a missing datapoint entered with a '?'
should do what you want. I've not tried this, but I think that does the
trick.
--
Kevin Oberman, Part time kid herder and retired Network Engineer
E-mail: rko...@gm...
PGP Fingerprint: D03FB98AFA78E3B78C1694B318AB39EF1B055683

Hello,

If you type "help missing" in an interactive gnuplot session, you'll see 
the following:

Gnuplot makes a distinction between missing data and
invalid data (e.g. "NaN", 1/0.).  For example invalid data causes a gap in a
line drawn through sequential data points; missing data does not.

So you should *not* set "x" as the string for missing data. Instead, try 
this:

$DATA << END
1 1
2 2
3 x
4 4
5 5
6 x
7 7
8 8
END

plot $DATA u 1:($2) w linespoints

You'll get a straight line with a two gaps.

I hope this helps.

Gabriel


Le 2021-09-20 à 18 h 45, Dave Horsfall a écrit :
> Anyone?
>
> -- Dave
>
> ---------- Forwarded message ----------
> Date: Tue, 31 Aug 2021 08:37:08 +1000 (EST)
> From: Dave Horsfall <da...@ho...>
> To: Gnuplot Mailing List <gnu...@li...>
> Subject: Gnuplot is interpolating between missing data
>
> Gnuplot (on a Mac), 5.4p2
>
> I have some missing data which I indicate with "x" (set datafile 
> missing "x"), yet Gnuplot seems to be drawing a straight line between 
> the valid data. Shouldn't it be blank instead?
>
> -- Dave
>
>
> _______________________________________________
> gnuplot-info mailing list
> gnu...@li...
> Membership management via: 
> https://lists.sourceforge.net/lists/listinfo/gnuplot-info

Anyone?

-- Dave

---------- Forwarded message ----------
Date: Tue, 31 Aug 2021 08:37:08 +1000 (EST)
From: Dave Horsfall <da...@ho...>
To: Gnuplot Mailing List <gnu...@li...>
Subject: Gnuplot is interpolating between missing data

Gnuplot (on a Mac), 5.4p2

I have some missing data which I indicate with "x" (set datafile missing "x"), 
yet Gnuplot seems to be drawing a straight line between the valid data. 
Shouldn't it be blank instead?

-- Dave