[R] mild and extreme outliers in boxplot

[Gabor Grothendieck]

On Thu, Aug 20, 2009 at 3:23 AM, Gavin Simpson<gavin.simpson at ucl.ac.uk> wrote:
> Hi Rolf,
>
> On Thu, 2009-08-20 at 12:31 +1200, Rolf Turner wrote:
>>
>> I despair.  Why do you keep insisting that black is white?
>> The OP wanted to be able to specify an argument to boxplot()
>> that would cause it to plot mild and extreme outliers with
>> different symbols.
>
> Then you and I were reading different emails Rolf. I quote:
> [Sorry I don't have the original email, so this from Gmane:
> http://article.gmane.org/gmane.comp.lang.r.general/159957 ]
>
>
>> dear all,
>>
>> could somebody tell me how I can plot mild outliers as a circle(°) and
>> extreme outliers as an asterisk(*) in a box-whisker plot?
>>
>> Thanks very much in advance
>
> The OP wants to use a different pch for the points that are plotted at
> the extremes of the distribution of points. It is up to him, as I
> pointed out, to define what "extreme" means in this context...
>
>>
>>       THIS CAN'T BE DONE!!!
>
> ...but the plotting he wanted to do, can be done! I showed one way it
> could be done. Why do you insist he can't do what he wants?

I think what he means is that outpch can only supply one symbol per
boxplot so to get the result you need to superimpose the points on the
boxplot using points() as in my post.  If you supply a vector to outpch
then successive values are used for successive boxplots, not for
successive outliers in the same boxplot.