[R] formatting output

[Marc Schwartz]

On Mon, 2005-02-28 at 14:05 -0400, Peyuco Porras Porras . wrote:
>    Dear R-users
> 
>    A  basic  question  that I wasn't able to solve: Is it possible to get
>    the  results  of the function 'quantile' expressed as data.frame? What
>    I'm  doing  is  to  apply the following code to get the quantiles in a
>    particular dataset:
> 
>    tmp<-tapply(data$DEN,list(Age=data$AGE,Sex=data$SEX),quantile)
> 
>    and  then I save this output to HTML using the library R2HTML. However
>    in  order  to  format  the  tables  in  HTML I have to use the command
>    HTML.data.frame(...)  which  allows  me  to  define,  for example, the
>    number of digits in the html table.
> 
>    But  the object 'tmp' is not a dataframe and I can't coarce it to this
>    format.  Is  it  possible  to  get  the  results of this function as a
>    dataframe? I  know  that  I'm probably missing some important concepts
>    here but Im not very good in programming.
> 
>    Any hint will be appreciated


Here is one approach, using an expansion of one of the examples in
?tapply:

# Get quantiles of 'breaks' for each combination of 'wool' and 'tension'
> my.tmp <- tapply(warpbreaks$breaks, 
                   list(warpbreaks$wool, warpbreaks$tension), quantile)

# Note that my.tmp is a 2 x 6 matrix of list elements
# with each list element being the results of quantile
# on each combination of 'wool' and 'tension'

> my.tmp
  L         M         H
A Numeric,5 Numeric,5 Numeric,5
B Numeric,5 Numeric,5 Numeric,5


For example:

> my.tmp[1, 1]
[[1]]
  0%  25%  50%  75% 100%
  25   26   51   54   70


Now get this into a manageable structure by taking each list element in
my.tmp and converting it into a row in a new matrix, use:

> my.mat <- do.call("rbind", my.tmp)

> my.mat
     0% 25% 50% 75% 100%
[1,] 25  26  51  54   70
[2,] 14  20  29  31   44
[3,] 12  18  21  30   36
[4,] 16  21  28  39   42
[5,] 10  18  24  28   43
[6,] 13  15  17  21   28


>From there, you can set the rownames for the matrix, as you require,
based upon the combinations of the vectors in your data. For example,
using expand.grid():

> my.names <- expand.grid(dimnames(my.tmp))

> my.names
  Var1 Var2
1    A    L
2    B    L
3    A    M
4    B    M
5    A    H
6    B    H

> rownames(my.mat) <- paste(my.names$Var1, my.names$Var2, sep = ":")

> my.mat
    0% 25% 50% 75% 100%
A:L 25  26  51  54   70
B:L 14  20  29  31   44
A:M 12  18  21  30   36
B:M 16  21  28  39   42
A:H 10  18  24  28   43
B:H 13  15  17  21   28


There might be an easier way, but that's my quick thought.

HTH,

Marc Schwartz