[R] lattice contourplots

[Jeff Newmiller]

Sorry, my first response was too hasty. Per the help file you need to use some kind of interpolation method to resample whatever random data points you have in a regular grid before you give it to contourplot.
---------------------------------------------------------------------------
Jeff Newmiller                        The     .....       .....  Go Live...
DCN:<jdnewmil at dcn.davis.ca.us>        Basics: ##.#.       ##.#.  Live Go...
                                      Live:   OO#.. Dead: OO#..  Playing
Research Engineer (Solar/Batteries            O.O#.       #.O#.  with
/Software/Embedded Controllers)               .OO#.       .OO#.  rocks...1k
--------------------------------------------------------------------------- 
Sent from my phone. Please excuse my brevity.

On June 1, 2015 12:22:12 PM EDT, David Winsemius <dwinsemius at comcast.net> wrote:
>
>On Jun 1, 2015, at 5:39 AM, ivo welch wrote:
>
>> thank you.  yes, I got bitten by "FAQ 7.22: Why do lattice/trellis
>> graphics not work?"  It had never occurred to me that this could be
>> expected behavior or a FAQ.  (didn't show up in a google search for
>> contourplot.)  Unless one knows, this is a puzzler.  Thanks for the
>> pointer.  so, a minimum working example is
>> 
>>    require(lattice)
>>    d <- data.frame( expand.grid( x = seq(0,6,length.out=100), y =
>> seq(0,6,length.out=100) ) )
>>    d <- within(d, z <- sin( (x+y) ))
>>    xx <- contourplot( z ~ x * y, data = d)
>>    print(xx)  ## necessary in source code, but not in interactive
>mode
>
>I do not think FAQ 7.22 was the problem. Look at the structure of your
>two different 'd' objects. One of them had the proper long form of an
>x-y grid with z values while the other one did not. (Also a Mac user
>but this has nothing to do with Macs.) I was getting a "blank plot"
>with your first version but getting the expected contours at minus 45
>degrees for the second version regardless of any print commands. But my
>"blank plot had axes that had an appropriate range but was empty inside
>(no contours) so it wasn't a completely empty plotting event. I don't
>think `contourplot` was able to make any sense out of the completely
>irregular data in the first instance. I do agree with Newmiller's
>second guess that the description of x (i.e.formula) on the help page
>is the correct reference, but I think you misinterpreted his meaning.