R help - Nov 1998 - Question about dotplot

Hello:

I am playing with the dotplot command (in preparation for 
teaching an EDA course). I do not understand the following 
behaviour of dotplot when x is a vector. These examples 
have been tried on both Linux version 0.63 and guido's 
Windows version 0.63 running under Windows 95.
I searched the bug reporting system, but found no mention
of any report concerning dotplot.

x <- 1:26
names(x) <- letters

# y has 10 zeros, 10 ones, and 6 twos.
y <- c(rep(0,10), rep(1,10), rep(2,6)) 
y <- factor(y, labels = c("G1","G2", "G3"))

# lcolor="black" since dotted gray lines don't display
# properly in my Windows version.

dotplot(x,groups=y, lcolor="black") 
--------------------------------------------------------
The dotplot groups the a-f components of x and labels them 
G1, groups the g-p components of x and labels them G2, and 
groups the q-z components of x and labels them G3. I would 
have expected the first 10 components of x to be grouped 
together with the G1 label, followed by the next 10 
components with label G2, then the last 6 components with 
label G3.
---------------------------------------------------------

# z has 5 zeros, then 10 ones, then 5 zeros, then 6 ones.
z <- c(rep(0,5), rep(1,10), rep(0,5), rep(1,6))
z <- factor(z, labels = c("G1", "G2"))
dotplot(x,groups=z, lcolor="black")
----------------------------------------------------------

This groups components a,b,c,d,e,p, then a break, then 
components q,r,s,t, then a break, then f to o, then a break, 
then u to z. The G1 label is aligned at the same level as 
the d entry, and the G2 label is aligned with the m entry.

Does this make sense?

Thank you.

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at
stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._

Norm Josephy <NJOSEPHY at bentley.edu> writes:
> This groups components a,b,c,d,e,p, then a break, then 
> components q,r,s,t, then a break, then f to o, then a break, 
> then u to z. The G1 label is aligned at the same level as 
> the d entry, and the G2 label is aligned with the m entry.
> 
> Does this make sense?
Happens on Unix too, and no: it doesn't make sense...

I believe the blunder is in here:
    else {
        o <- rev(order(as.numeric(groups)))
        x <- x[o]
        groups <- groups[o]
        offset <- cumsum(c(0, diff(as.numeric(groups)[o]) != 
            0))
        y <- 1:n + 2 * offset
        ylim <- range(0, y + 2)
    }

groups is getting ordered *twice* there. Dropping the [o] in the
diff() makes the extra gaps go away, but now the group labels look
wrong. Oh, well.

-- 
   O__  ---- Peter Dalgaard             Blegdamsvej 3  
  c/ /'_ --- Dept. of Biostatistics     2200 Cph. N   
 (*) \(*) -- University of Copenhagen   Denmark      Ph: (+45) 35327918
~~~~~~~~~~ - (p.dalgaard at biostat.ku.dk)             FAX: (+45) 35327907
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at
stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._