Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.graphics.apps.gnuplot > #3871 > unrolled thread

For loop not working as I expected

Back to article view | Back to comp.graphics.apps.gnuplot

Contents

  For loop not working as I expected Chris Elvidge <chris@mshome.net> - 2018-02-17 16:46 +0000
    Re: For loop not working as I expected Karl Ratzsch <mail.kfr@gmx.net> - 2018-02-19 08:15 +0100
      Re: For loop not working as I expected - but now solved Chris Elvidge <chris@mshome.net> - 2018-02-19 12:50 +0000
        Re: For loop not working as I expected - but now solved Karl Ratzsch <mail.kfr@gmx.net> - 2018-02-19 14:26 +0100
          Re: For loop not working as I expected - but now solved Chris Elvidge <chris@mshome.net> - 2018-02-19 14:22 +0000
            Re: For loop not working as I expected - but now solved Karl Ratzsch <mail.kfr@gmx.net> - 2018-02-19 16:11 +0100
              Re: For loop not working as I expected - but now solved Chris Elvidge <chris@mshome.net> - 2018-02-19 15:25 +0000
          Re: For loop not working as I expected - but now solved Ethan A Merritt <EAMerritt@gmail.com> - 2018-02-21 22:38 -0800
            Re: For loop not working as I expected - but now solved Karl Ratzsch <mail.kfr@gmx.net> - 2018-02-23 18:50 +0100

#3871 — For loop not working as I expected

Hi all,

I'm a relative newcomer to gnuplot and can't find meaningful answers on 
google. Gnuplot Version 5.0 patchlevel 5

This is a section of my gnuplot file:

F(n) = system("date -d\"-".sprintf('%d',n)." days\" 
+\"/var/tmp/cpu-gpu-temp_%Y%m%d.dat\"")

#F(n) = system("date -d\"-".n." days\" 
+\"/var/tmp/cpu-gpu-temp_%Y%m%d.dat\"")

do for [i=1:7] { print F(8-i); }

plot [j=1:7] F(8-j) u 1:4 w l lw 2 lc rgb "orange" t "CPU", '' u 1:5 w l 
lw 2 lc rgb "green" t "GPU"

#plot [j=1:7] F(sprintf('%d',8-j)) u 1:4 w l lw 2 lc rgb "orange" t 
"CPU", '' u 1:5 w l lw 2 lc rgb "green" t "GPU"

The print command gives the correct output whichever F(n) definition I use:
/var/tmp/cpu-gpu-temp_20180210.dat
/var/tmp/cpu-gpu-temp_20180211.dat
/var/tmp/cpu-gpu-temp_20180212.dat
/var/tmp/cpu-gpu-temp_20180213.dat
/var/tmp/cpu-gpu-temp_20180214.dat
/var/tmp/cpu-gpu-temp_20180215.dat
/var/tmp/cpu-gpu-temp_20180216.dat


but the plot command gives on of two errors:
either:
warning: encountered a string when expecting a number
Did you try to generate a file name using dummy variable x or y?

or:
internal error : STRING operator applied to non-STRING type

I've worked round it, but wondered where I am making mistakes in the 
implied for loop in the plot command.

As an aside, feeding the file list from bash as $filelist, is there any 
way of generating ARGn to be used as input file?

e.g. plot [i=1:ARGC] ARGi u 1:5 etc? All I get is ARGi is unknown/not a 
file.

Thanks for any input.
Cheers


-- 

Chris Elvidge, England

[toc] | [next] | [standalone]

#3874

Am 17.02.2018 um 17:46 schrieb Chris Elvidge:

> plot [j=1:7] F(8-j) u 1:4 w l lw 2 lc rgb "orange" t "CPU", '' u 1:5 w l 
> lw 2 lc rgb "green" t "GPU"

The second part of

   plot [i=1:4] fname(i) us 1:2, '' us 1:2

cannot work. '' just contains the last file name, in verbatim, that was
in the plot command. It doesn't expect a function, and especially cannot
contain the whole iteration. The error message could be a bit clearer,
but obviously nobody thought of the specific misuse you tried. ;-)

You have to repeat the iteration :

 plot [i=1:4] fname(i) us 1:2 w lc i, for [i=1:4] fname(i) us 1:2 w lc i


Best, Karl

[toc] | [prev] | [next] | [standalone]

#3875 — Re: For loop not working as I expected - but now solved

On 19/02/2018 07:15 am, Karl Ratzsch wrote:
> Am 17.02.2018 um 17:46 schrieb Chris Elvidge:
> 
>> plot [j=1:7] F(8-j) u 1:4 w l lw 2 lc rgb "orange" t "CPU", '' u 1:5 w l
>> lw 2 lc rgb "green" t "GPU"
> 
> The second part of
> 
>     plot [i=1:4] fname(i) us 1:2, '' us 1:2
> 
> cannot work. '' just contains the last file name, in verbatim, that was
> in the plot command. It doesn't expect a function, and especially cannot
> contain the whole iteration. The error message could be a bit clearer,
> but obviously nobody thought of the specific misuse you tried. ;-)
> 
> You have to repeat the iteration :
> 
>   plot [i=1:4] fname(i) us 1:2 w lc i, for [i=1:4] fname(i) us 1:2 w lc i
> 
> 
> Best, Karl
> 

Sorry Karl, that didn't work either. But thanks for pointing out the 
need for the second "for [i=1:4]" bit.

However, this did work:

F(n) = system("date -d\"-".n." days\" 
\"+/var/tmp/cpu-gpu-temp_%Y%m%d.dat\"")
plot for [j=1:7] F(8-j) u 1:4 w l lw 2 lc rgb "orange" noti, for [j=1:7] 
F(8-j) u 1:5 w l lw 2 lc rgb "web-green" noti

Note the extra "for" after the plot command. Could this be a bug in the 
manual?

Strangely, with "plot for" and '' as the second file specifier, I get 
all the files for the first plot, but only the last file for the second 
plot.

Cheers

-- 

Chris Elvidge, England

[toc] | [prev] | [next] | [standalone]

#3876 — Re: For loop not working as I expected - but now solved

Am 19.02.2018 um 13:50 schrieb Chris Elvidge:

> plot for [j=1:7] F(8-j) u 1:4 w l lw 2 lc rgb "orange" noti, for [j=1:7] 
> F(8-j) u 1:5 w l lw 2 lc rgb "web-green" noti
> 
> Note the extra "for" after the plot command. Could this be a bug in the 
> manual?

Sorry, I totally overlooked that! Of course it needs to be "for [:]" in
both the first and the second subplot. Where does the manual say
otherwise? That would be an error and you should file a bug report.

The command

  plot [j=1:7] dataf ...

should assign the value 1 to "j", and make a plot with abscissa
range[1:7]. (The assignment is not actually done, but the parser also
doesn't throw an error. It arguably should do one or the other, so
that's a small bug.)


> Strangely, with "plot for" and '' as the second file specifier, I get 
> all the files for the first plot, but only the last file for the second 
> plot.

That's the expected behaviour. The special filename '' just contains the
last file(name) that was plotted, not the whole iteration.


  Karl

[toc] | [prev] | [next] | [standalone]

#3877 — Re: For loop not working as I expected - but now solved

On 19/02/2018 01:26 pm, Karl Ratzsch wrote:
> Am 19.02.2018 um 13:50 schrieb Chris Elvidge:
> 
>> plot for [j=1:7] F(8-j) u 1:4 w l lw 2 lc rgb "orange" noti, for [j=1:7]
>> F(8-j) u 1:5 w l lw 2 lc rgb "web-green" noti
>>
>> Note the extra "for" after the plot command. Could this be a bug in the
>> manual?
> 
> Sorry, I totally overlooked that! Of course it needs to be "for [:]" in
> both the first and the second subplot. Where does the manual say
> otherwise? That would be an error and you should file a bug report.
> 
> The command
> 
>    plot [j=1:7] dataf ...
> 
> should assign the value 1 to "j", and make a plot with abscissa
> range[1:7]. (The assignment is not actually done, but the parser also
> doesn't throw an error. It arguably should do one or the other, so
> that's a small bug.)

I put it down to my misreading (i.e. not fully understanding what is 
going on) of this (first example) where 'plot [-10:10]' is used; 'for' 
is not mentioned:

http://gnuplot.sourceforge.net/demo_svg_5.0/simple.html

Anyhow, thanks for the pointers. With any luck my skills will improve.

Cheers

-- 

Chris Elvidge, England

[toc] | [prev] | [next] | [standalone]

#3878 — Re: For loop not working as I expected - but now solved

Am 19.02.2018 um 15:22 schrieb Chris Elvidge:

> I put it down to my misreading (i.e. not fully understanding what is 
> going on) of this (first example) where 'plot [-10:10]' is used; 'for' 
> is not mentioned:
> 
> http://gnuplot.sourceforge.net/demo_svg_5.0/simple.html


The full syntax for plot iteration is this:

  plot {[xmin:xmax] {[ymin:ymax]}} for [i=a:b] F(x)

, where everything in curly brackets is optional. Try

  plot [-2:2] [-5:10] x**2

against

  plot for [i=0:3] x**i

xmin,xmax,ymin,ymax just give the length of the axes, while iteration
needs the "for" keyword.

[toc] | [prev] | [next] | [standalone]

#3879 — Re: For loop not working as I expected - but now solved

On 19/02/2018 03:11 pm, Karl Ratzsch wrote:
> Am 19.02.2018 um 15:22 schrieb Chris Elvidge:
> 
>> I put it down to my misreading (i.e. not fully understanding what is
>> going on) of this (first example) where 'plot [-10:10]' is used; 'for'
>> is not mentioned:
>>
>> http://gnuplot.sourceforge.net/demo_svg_5.0/simple.html
> 
> 
> The full syntax for plot iteration is this:
> 
>    plot {[xmin:xmax] {[ymin:ymax]}} for [i=a:b] F(x)
> 
> , where everything in curly brackets is optional. Try
> 
>    plot [-2:2] [-5:10] x**2
> 
> against
> 
>    plot for [i=0:3] x**i
> 
> xmin,xmax,ymin,ymax just give the length of the axes, while iteration
> needs the "for" keyword.
> 
Got it now - I think!
Cheers again


-- 

Chris Elvidge, England

[toc] | [prev] | [next] | [standalone]

#3884 — Re: For loop not working as I expected - but now solved

Karl Ratzsch wrote:


> The command
> 
>   plot [j=1:7] dataf ...
> 
> should assign the value 1 to "j", and make a plot with abscissa
> range[1:7]. (The assignment is not actually done, but the parser also
> doesn't throw an error. It arguably should do one or the other, so
> that's a small bug.)
 
No bug there.  The dependent variable becomes j rather than x,
allowing you to say:

    plot [theta=1:7] sin(theta)


	Ethan

[toc] | [prev] | [next] | [standalone]

#3888 — Re: For loop not working as I expected - but now solved

Am 22.02.2018 um 07:38 schrieb Ethan A Merritt:
> Karl Ratzsch wrote:
> 
> 
>> The command
>>
>>   plot [j=1:7] dataf ...
>>
>> should assign the value 1 to "j", and make a plot with abscissa
>> range[1:7]. (The assignment is not actually done, but the parser also
>> doesn't throw an error. It arguably should do one or the other, so
>> that's a small bug.)
>  
> No bug there.  The dependent variable becomes j rather than x,
> allowing you to say:
> 
>     plot [theta=1:7] sin(theta)

Yes, of course. Sorry for the noise.

  Karl

[toc] | [prev] | [standalone]

Back to top | Article view | comp.graphics.apps.gnuplot

csiph-web