stem plot problem

[Ben Abbott]

On Oct 17, 2008, at 9:18 PM, Ben Abbott wrote:

>
> On Oct 17, 2008, at 5:51 PM, Ben Abbott wrote:
>
>> On Friday, October 17, 2008, at 02:45PM, "John W. Eaton" <jwe at bevo.che.wisc.edu 
>> > wrote:
>>> I'm seeing the following problem with the current development  
>>> sources
>>> (this is demo 7 in stem.m):
>>>
>>> octave:1>  x = [0:10].';
>>> octave:2>  y = [sin(x), cos(x)];
>>> octave:3>  h = stem (x, y);
>>> octave:4>  set (h(2), "color", "g");
>>> octave:5>  set (h(1), "basevalue", 1)
>>> error: number of rows must match (33 != 11) near line 468, column 33
>>> error: called from:
>>> error:   /home/jwe/src/octave/scripts/plot/__go_draw_axes__.m at  
>>> line 468, column 24
>>> error:   /home/jwe/src/octave/scripts/plot/__go_draw_figure__.m at  
>>> line 56, column 8
>>> error:   /home/jwe/src/octave/scripts/plot/gnuplot_drawnow.m at  
>>> line 68, column 5
>>>
>>> The problem appears to be this concatenation at line 465 in
>>> __go_draw_axes__.m:
>>>
>>> 	      data{data_idx} = [xdat, ydat]';
>>>
>>> Again, I don't understand what is supposed to be happening, or why
>>> xdat has 11 elements and ydat has 33 at this point.  I could use
>>> some help fixing this problem.
>>>
>>> jwe
>>
>> I don't think the problem is in __go_draw_axes__.m
>>
>> If the last line of that demo is removed (the one below), there is  
>> no error.
>>
>> %! set (h(1), "basevalue", -1)
>>
>> I inserted the lines below ahead of the error
>>
>> %! get(get(get (h(1),'children')(1), 'children'), 'xdata')
>> %! get(get(get (h(1),'children')(1), 'children'), 'ydata')
>> %! get(get(get (h(1),'children')(2), 'children'), 'xdata')
>> %! get(get(get (h(1),'children')(2), 'children'), 'ydata')
>>
>> The 1st & 2nd produce a 1x11 vector
>> The 3rd & 4th produce a 1x33 vector
>>
>> After examining the line that fails in __go_draw_axes__.m, it  
>> appears an attempt is made to plot the xdata from the 1st with the  
>> ydata from the 4th.
>>
>> I'll have to bone up on what "basevalue" is before going any deeper.
>>
>>
>> The handle h(1) points to an hggroup. Looking at the lines of the  
>> group all the xdata and ydata appear to be correct. Once the set  
>> command gets run, I'm not sure what the sequence of events is.
>>
>> Ben
>
> ok, I'm in over my head, but it appears that the problem is with  
> "update_props" .... and for those "in the know", the question is  
> which "update_props"?
>
> In __stem__.m it is defined as ...
>
> 	509 function update_props (h, d)
> 	510   kids = get (h, "children");
> 	511   set (kids(1), "color", get (h, "color"),
> 	512        "linewidth", get (h, "linewidth"),
> 	513        "linestyle", get (h, "linestyle"));
> 	514   set (kids(2), "color", get (h, "color"),
> 	515        "marker", get (h, "marker"),
> 	516        "markerfacecolor", get (h, "markerfacecolor"),
> 	517        "markersize", get (h, "markersize"));
> 	518 endfunction
>
> In __add_line_series__.m
>
> 	 52 function update_props (h, d)
> 	 53   set (get (h, "children"), "color", get (h, "color"),
> 	 54        "linewidth", get (h, "linewidth"),
> 	 55        "linestyle", get (h, "linestyle"),
> 	 56        "marker", get (h, "marker"),
> 	 57        "markerfacecolor", get (h, "markerfacecolor"),
> 	 58        "markeredgecolor", get (h, "markeredgecolor"),
> 	 59        "markersize", get (h, "markersize"),
> 	 60        "xdata", get (h, "xdata"),
> 	 61        "ydata", get (h, "ydata"),
> 	 62        "zdata", get (h, "zdata"));
> 	 63 endfunction
>
> I suggest such after pasting the lines below into the octave console
>
> 	 x = [0 : 10].';
> 	 y = [sin(x), cos(x)];
> 	 h = stem (x, y);
> 	 set (h(2), "color", "g");
> 	 set (h(1), "basevalue", -1)
>
> Which produced the following ...
>
> octave:158>  x = [0 : 10].';
> octave:159>  y = [sin(x), cos(x)];
> octave:160>  h = stem (x, y);
> octave:161>  set (h(2), "color", "g");
> octave:162>  set (h(1), "basevalue", -1)
>
> error: invalid value for array property "ydata"
> error: set: expecting argument 18 to be a property name
> error: called from:
> error:   /Users/bpabbott/Development/mercurial/octave-3.1.51/scripts/ 
> plot/__add_line_series__.m at line 56, column 1
> error: invalid graphics handle
> error:   /Users/bpabbott/Development/mercurial/octave-3.1.51/scripts/ 
> plot/__stem__.m at line 509, column 3
> error: invalid graphics handle
> error:   /Users/bpabbott/Development/mercurial/octave-3.1.51/scripts/ 
> plot/__stem__.m at line 479, column 2
> error: invalid graphics handle
> error:   /Users/bpabbott/Development/mercurial/octave-3.1.51/scripts/ 
> plot/__stem__.m at line 495, column 5
> octave:162>
>
> This is getting a bit too much for a guy on a Friday night with 3  
> beers behind him, so I'm off to other entertainment .... until  
> tomorrow ;-)
>
> Ben

ok a new observation that even a guy with four beers behind him can see

	 x = [0 : 10].';
	 y = [sin(x), cos(x)];
	 h = stem (x, y);
	 set (h(2), "color", "g");
	 set (h(1), "basevalue", -1)

The last command produces the error, but if move_baeeline() in  
__stem__.m as

487 function move_baseline (h, d)
488   b0 = get (h, "basevalue");
489   bl = get (h, "baseline");
490
491   if (get (bl, "ydata") != [b0, b0])
492     set (bl, "ydata", [b0, b0]);
493   endif
494
495   kids = get (h, "children");
496   yt = get(h, "ydata");
497   if (size (yt, 1) == 1)
498     yt = yt(:)';
499     ny = length (yt);
500     yt = [b0 * ones(1, ny); yt; NaN(1, ny)](:);
501   else
502     yt = yt(:);
503     ny = length (yt);
504     yt = [b0 * ones(ny, 1), yt, NaN(ny, 1)];
505   endif
506   set (kids(1), "ydata", yt);
507 endfunction

and the command below is submitted twice

	 set (h(1), "basevalue", -1)

All works as it should.

Perhaps there is something of a recursive nature that is breaking this?

Ben