abbreviate or wrap dimname labels

Fri, Apr 15, 2005 10:29 AM

On Fri, 2005-04-15 at 12:12 -0400, Michael Friendly wrote:

For a variety of displays (mosaicplots, barplots, ...)
one often wants to either abbreviate or wrap long labels,
particularly when these are made up of several words.
In general, it would be nice to have a function,

abbreviate.or.wrap <-
    function(x, maxlength=10, maxlines=2, split=" ") {
}

that would take a character vector or a list of vectors, x,
and try to abbreviate or wrap them to fit approximately
the maxlength and maxlines constraints, using the split
argument to specify allowable characters to wrap to multiple
lines.

For example, this two-way table has dimnames too long to
be displayed nicely in a mosaicplot:

 > library(catspec)
 > library(vcd)
 >
 > data(FHtab)
 > FHtab<-as.data.frame(FHtab)
 >
 > xtable <- xtabs(Freq ~ .,FHtab)
 > lab <- dimnames(xtable)
 > lab

$OccFather
[1] "Upper nonmanual" "Lower nonmanual" "Upper manual"    "Lower manual"
[5] "Farm"

$OccSon
[1] "Upper nonmanual" "Lower nonmanual" "Upper manual"    "Lower manual"
[5] "Farm"

abbreviate works here, but gives results that aren't very readable:

 > lapply(lab, abbreviate, 8)

$OccFather
Upper nonmanual Lower nonmanual    Upper manual    Lower manual       Farm
      "Upprnnmn"      "Lwrnnmnl"      "Uppermnl"      "Lowermnl" 
   "Farm"

$OccSon
Upper nonmanual Lower nonmanual    Upper manual    Lower manual 
    Farm
      "Upprnnmn"      "Lwrnnmnl"      "Uppermnl"      "Lowermnl" 
   "Farm"

In a related thread, Marc Schwartz proposed a solution for wrapping
labels, based on

 >short.labels <- sapply(labels, function(x) paste(strwrap(x,

                          10), collapse = "\n"), USE.NAMES = FALSE)

But, my attempt to use strwrap in my context gives a single string
for each set of dimension names:

 > stack.lab <-function(x) { paste(strwrap(x,10), collapse = "\n") }
 > lapply(lab, stack.lab)

$OccFather
[1] "Upper\nnonmanual\nLower\nnonmanual\nUpper\nmanual\nLower\nmanual\nFarm"

$OccSon
[1] "Upper\nnonmanual\nLower\nnonmanual\nUpper\nmanual\nLower\nmanual\nFarm"

For my particular example, I can do what I want with gsub, but it is
hardly general:

 > lab[[1]] <- gsub(" ","\n", lab[[1]])
 > lab[[2]] <- lab[[1]]   # cheating: I know it's a square table
 > lab

$OccFather
[1] "Upper\nnonmanual" "Lower\nnonmanual" "Upper\nmanual" 
"Lower\nmanual"
[5] "Farm"

$OccSon
[1] "Upper\nnonmanual" "Lower\nnonmanual" "Upper\nmanual" 
"Lower\nmanual"
[5] "Farm"

 > dimnames(xtable) <- lab

Then,
mosaicplot(xtable, shade=TRUE)
gives a nice display!

Can anyone help with a more general solution for wrapping labels
or abbreviate.or.wrap()?

thanks,
-Michael

Michael,

This is not completely generic (I have not used abbreviate() here) and
it could take some further fine tuning and perhaps even consideration of
creating a generic method. However, a possible solution to the problem
of using my previous approach on a list object and giving some
flexibility to also handle vectors:


# Core wrapping function
wrap.it <- function(x, len)
{ 
  sapply(x, function(y) paste(strwrap(y, len), 
                        collapse = "\n"), 
         USE.NAMES = FALSE)
}


# Call this function with a list or vector
wrap.labels <- function(x, len)
{
  if (is.list(x))
  {
    lapply(x, wrap.it, len)
  } else {
    wrap.it(x, len)
  }
}



Thus, for your labels in a list:

$OccFather
[1] "Upper\nnonmanual" "Lower\nnonmanual" "Upper\nmanual"   
[4] "Lower\nmanual"    "Farm"            

$OccSon
[1] "Upper\nnonmanual" "Lower\nnonmanual" "Upper\nmanual"   
[4] "Lower\nmanual"    "Farm"   


and for the example vector in my prior post:

[1] "This is\na long\nlabel 1"  "This is\na long\nlabel 2" 
 [3] "This is\na long\nlabel 3"  "This is\na long\nlabel 4" 
 [5] "This is\na long\nlabel 5"  "This is\na long\nlabel 6" 
 [7] "This is\na long\nlabel 7"  "This is\na long\nlabel 8" 
 [9] "This is\na long\nlabel 9"  "This is\na long\nlabel 10"


To incorporate abbreviate() here, you could perhaps modify the
wrap.labels() syntax to use a "wrap = TRUE/FALSE" argument to explicitly
indicate which approach you want, or perhaps develop some decision tree
approach to automate the process.

HTH,

Marc Schwartz

abbreviate or wrap dimname labels

Thread (6 messages)