Wow, this works, Gabor, but I am mystified. I would have tought an  
expression needed the word expression, and/or a text string needed to  
be within quotes. What is happening here, exactly? Why the use of  
"~"? I tried without and it no longer works.

Thanks in advance,

Denis
Le 07-03-18 à 08:59, Gabor Grothendieck a écrit :

> Sorry, legend= was omitted:
>
> plot(1:10)
> legend("topleft", legend = This ~ study ~ italic(n) == 3293)
>
> On 3/18/07, Chabot Denis <[EMAIL PROTECTED]> wrote:
>> Thank you Marc, Jim and Gabor,
>>
>> I like the solution with "expression", nice and simple. Gabor, your
>> solution did not work, probably just a matter of putting the text
>> inside an expression?
>>
>> However it would be nice if the help system pointed to it. A search
>> on "italics" brought me nothing, one on "italic" gave me 4 hits, none
>> useful. And reading the help on plotmath, I found no mention of  
>> italic
>> (). Where can we suggest additions to the help system?
>>
>> I must plead guilty to have forgotten a RSiteSearch before posting. I
>> just did and I think I might have figured out something out there.
>> But your answers were nice and to the point!
>>
>> Cheers,
>>
>> Denis
>> Le 07-03-17 à 23:30, Marc Schwartz a écrit :
>>
>> > On Sat, 2007-03-17 at 21:56 -0500, Marc Schwartz wrote:
>> >> On Sat, 2007-03-17 at 22:01 -0400, Chabot Denis wrote:
>> >>> Hi,
>> >>>
>> >>> As part of the legend to a plot, I need to have the "n" in  
>> italics
>> >>> because it is a requirement of the journal I aim to publish in:
>> >>> "This study, n = 3293"
>> >>>
>> >>> Presently I have:
>> >>> legend(20, 105, "This study, n = 3293", pch=1,  col=rgb 
>> (0,0,0,0.5),
>> >>>                  pt.cex=0.3, cex=0.8, bty="n")
>> >>>
>> >>> I suppose I could leave a blank in place of the "n", then issue a
>> >>> text call where I'd use font=3 for a single letter, "n". But  
>> it will
>> >>> be tricky to find the exact location to use.
>> >>>
>> >>> Is there a way to switch to font=3 just for one letter within a
>> >>> string?
>> >>>
>> >>> Thanks in advance,
>> >>>
>> >>> Denis Chabot
>> >>
>> >> Denis,
>> >>
>> >> Try something like this:
>> >>
>> >> plot(20, 100)
>> >>
>> >> leg <- legend(20, 105, "This study,    = 3293", pch = 1,
>> >>               col=rgb(0,0,0,0.5), pt.cex = 0.3, cex = 0.8,
>> >>               bty = "n")
>> >>
>> >> text(leg$text$x + strwidth("This study, ", cex = 0.8),
>> >>      leg$text$y, "n", font = 3, cex = 0.8, adj = c(0, 0.5))
>> >>
>> >>
>> >> Note that legend returns a list structure, which contains the x  
>> and y
>> >> coordinates of the start of the text strings that are plotted. So
>> >> I get
>> >> that information for your line of text.
>> >>
>> >> Next, I use strwidth() to calculate, in user coordinates, the
>> >> length of
>> >> the characters preceding the 'n', including spaces.  We add that
>> >> distance to the x coordinate returned in the legend call.
>> >>
>> >> I also use the 'adj' argument in the text() call, so that it is in
>> >> synch
>> >> with the same parameters in legend() for alignment with the other
>> >> letters.
>> >>
>> >> See ?strwidth for more information.
>> >>
>> >> You may have to tweak the horizontal spacing of the 'n' a bit,
>> >> depending
>> >> upon the rest of your graph.
>> >
>> > Denis,
>> >
>> > I thought of another approach, using plotmath.
>> >
>> > First, create a text expression, specifying that the 'n' should be
>> > italicized. Then use that expression in the legend() call.
>> >
>> > txt <- expression(paste("This study, ", italic(n), " = 3293"))
>> >
>> > plot(20, 100)
>> >
>> > legend(20, 105, txt, pch = 1, col=rgb(0,0,0,0.5),
>> >        pt.cex = 0.3, cex = 0.8, bty = "n")
>> >
>> >
>> > That's easier that the first solution.  See ?plotmath
>> >
>> > HTH,
>> >
>> > Marc Schwartz
>> >
>> >
>>
>>

______________________________________________
R-help@stat.math.ethz.ch mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.

Reply via email to