Re: splot and #+PLOT keyword

[Nick Dokos]

Eric S Fraga <e.fraga@ucl.ac.uk> writes:

> Hello,
>
> a question/problem regarding the #+plot: keyword.
>
> I am trying to plot out some data where the table looks like this:
>
> #+begin_src org
>   ,#+plot: ind:(6 7) deps:(1) with:"linespoints pt 7" set:"logscale xy" 
> type:3d
>   |     1 |   81 |      5 |      0 |   2.27 |     0.9729848950975623 |   
> 0.019370016994566613 | 0.0 |
>   |     2 |  179 |     12 |      2 |   3.68 |    0.42919355355596267 |      
> 0.098179980500945 | 0.0 |
>   |     3 |  192 |      6 |      4 |   3.73 |   0.022272788298562045 |     
> 107.57399021086516 | 0.0 |
>   |     4 |  207 |      7 |      8 |   3.78 |      5.793210638997738 |  
> 0.0012069891001225872 | 0.0 |
>   |     5 |  194 |     10 |     12 |   3.83 |    0.06356594000544429 |    
> 0.04724965431965522 | 0.0 |
>   |     6 |  216 |      8 |     14 |   3.89 |    0.06342050747033937 |  
> 0.0030380306687021346 | 0.0 |
>   |     7 |  193 |      7 |     18 |   3.94 |     0.0021538841210584 |    
> 0.05771587421360767 | 0.0 |
>   |     8 |  193 |      7 |     19 |   3.99 |  0.0010739216097561438 |    
> 0.10625133051680691 | 0.0 |
>   |     9 |  182 |      6 |     19 |   4.04 |  0.0014893478573963876 |    
> 0.03593357278451856 | 0.0 |
>   |    10 |  192 |      6 |     21 |   4.10 |   0.013251328328567616 |  
> 0.0006605631984014402 | 0.0 |
> #+end_src
>
>
> I have tried a variety of directives for the #+PLOT: keyword but none
> give me what I want.  I would like to plot column 1 (dependent variable)
> versus columns 6 and 7 (independent variables) in 3d.  (don't ask ;-))
>
> I have also tried:
>
>    #+plot: ind:6 deps:(7 1) ...
>
> but in all cases I seem to be getting somewhat random data plotted,
> possibly columns 2 versus 3 and 4, which makes no sense except that
> maybe the "|" table column separator is being retained in the data file
> created.  Is there any way to stop the data file from being deleted
> after plotting?  I can inspect the *gnuplot* buffer but cannot see the
> actual data.
>
> 2d plots work just fine, by the way.
>
> thank you,
> eric

Have you tried looking at the produced gnuplot script? It goes in a temp file
so it's a bit of a pain, but that's my fallback method when I'm really confused 
:-)

-- 
Nick

"There are only two hard problems in computer science: cache
invalidation, naming things, and off-by-one errors." -Martin Fowler