R help - Mar 2007 - objects of class "matrix" and mode "list"?

Hello everyone,

I cannot seem to find information about objects of class "matrix" and
mode
"list", and how to handle them (apart from flattening the list). I get
this
type of object from using sapply(). Sorry for the long example, but the code
below illustrates how I get this type of object. Is anyone aware of
documentation regarding this object?

Thanks very much,

Stephen

===== begin example ===
# I am just making up a fake data set
df <- data.frame(Day=rep(1:3,each=24),Hour=rep(1:24,times=3),
                 Name1=rnorm(24*3),Name2=rnorm(24*3))

# define a function to get a set of descriptive statistics
tmp <- function(x) {
  # this function will accept a data frame
  # and return a 1-row data frame of
  # max value, colname of max, min value, and colname of min
  return(data.frame(maxval=max(apply(x,2,max)),
                    maxloc=names(x)[which.max(apply(x,2,max))],
                    minval=min(apply(x,2,min)),
                    minloc=names(x)[which.min(apply(x,2,min))]))
}

# Now applying function to data:
# (1) split the data table by Day with split()
# (2) apply the tmp function defined above to each data frame from (1)
#     using lapply()
# (3) transpose the final matrix and convert it to a data frame
#     with mixed characters and numbers
#     using as.data.frame(), lapply(), and type.convert()
> final <- as.data.frame(lapply(as.data.frame(t(sapply(split(df[,-c(1:2)],
+                                                           
f=df$Day),tmp))),
+                               type.convert,as.is=TRUE))
Error in type.convert(x, na.strings, as.is, dec) : 
	the first argument must be of mode character

I thought sapply() would give me a data frame or matrix, which I would
transpose into a character matrix, to which I can apply type.convert()
and get the same matrix as what I would get from these two lines (Fold
function taken from Gabor's post on R-help a few years ago):

Fold <- function(f, x, L) for(e in L) x <- f(x, e)
final2 <- Fold(rbind,vector(),lapply(split(df[,-c(1:2)],f=day),tmp))
> print(c(class(final2),mode(final2)))
[1] "data.frame" "list"  

===================================================However, by my original
method, sapply() gives me a matrix with mode, "list"

intermediate1 <-
sapply(split(df[,-c(1:2)],f=df$Day),tmp)
> print(c(class(intermediate1),mode(intermediate1)))
[1] "matrix" "list"  

Transposing, still a matrix with mode list, not character:

intermediate2 <-
t(sapply(split(df[,-c(1:2)],f=day),tmp))
> print(c(class(intermediate2),mode(intermediate2)))
[1] "matrix" "list"  

Unclassing gives me the same thing...
> print(c(class(unclass(intermediate2)),mode(unclass(intermediate2))))
[1] "matrix" "list"  




 
____________________________________________________________________________________
Be a PS3 game guru.

It may help to (re-)read ?sapply a bit more in detail.  Simplification
is done only if it's "possible", and what "possible"
means is defined
there.

A list is a vector whose elements can be different objects, but a vector
nonetheless.  Thus a list can have dimensions.  E.g.,

R> a <- list(1, 1:2, 3, c("abc", "def"))
R> dim(a) <- c(2, 2)
R> a
     [,1]      [,2]       
[1,] 1         3          
[2,] Integer,2 Character,2

That sometimes can be extremely useful (not like the example above!).

Andy 

From: Stephen Tucker
> 
> Hello everyone,
> 
> I cannot seem to find information about objects of class 
> "matrix" and mode
> "list", and how to handle them (apart from flattening the 
> list). I get this
> type of object from using sapply(). Sorry for the long 
> example, but the code
> below illustrates how I get this type of object. Is anyone aware of
> documentation regarding this object?
> 
> Thanks very much,
> 
> Stephen
> 
> ===== begin example ===> 
> # I am just making up a fake data set
> df <- data.frame(Day=rep(1:3,each=24),Hour=rep(1:24,times=3),
>                  Name1=rnorm(24*3),Name2=rnorm(24*3))
> 
> # define a function to get a set of descriptive statistics
> tmp <- function(x) {
>   # this function will accept a data frame
>   # and return a 1-row data frame of
>   # max value, colname of max, min value, and colname of min
>   return(data.frame(maxval=max(apply(x,2,max)),
>                     maxloc=names(x)[which.max(apply(x,2,max))],
>                     minval=min(apply(x,2,min)),
>                     minloc=names(x)[which.min(apply(x,2,min))]))
> }
> 
> # Now applying function to data:
> # (1) split the data table by Day with split()
> # (2) apply the tmp function defined above to each data frame from (1)
> #     using lapply()
> # (3) transpose the final matrix and convert it to a data frame
> #     with mixed characters and numbers
> #     using as.data.frame(), lapply(), and type.convert()
> 
> > final <- 
> as.data.frame(lapply(as.data.frame(t(sapply(split(df[,-c(1:2)],
> +                                                           
> f=df$Day),tmp))),
> +                               type.convert,as.is=TRUE))
> Error in type.convert(x, na.strings, as.is, dec) : 
> 	the first argument must be of mode character
> 
> I thought sapply() would give me a data frame or matrix, which I would
> transpose into a character matrix, to which I can apply type.convert()
> and get the same matrix as what I would get from these two lines (Fold
> function taken from Gabor's post on R-help a few years ago):
> 
> Fold <- function(f, x, L) for(e in L) x <- f(x, e)
> final2 <- Fold(rbind,vector(),lapply(split(df[,-c(1:2)],f=day),tmp))
> 
> > print(c(class(final2),mode(final2)))
> [1] "data.frame" "list"  
> 
> ===================================================> However, by my
original method, sapply() gives me a matrix
> with mode, "list"
> 
> intermediate1 <- sapply(split(df[,-c(1:2)],f=df$Day),tmp)
> > print(c(class(intermediate1),mode(intermediate1)))
> [1] "matrix" "list"  
> 
> Transposing, still a matrix with mode list, not character:
> 
> intermediate2 <- t(sapply(split(df[,-c(1:2)],f=day),tmp))
> > print(c(class(intermediate2),mode(intermediate2)))
> [1] "matrix" "list"  
> 
> Unclassing gives me the same thing...
> 
> > print(c(class(unclass(intermediate2)),mode(unclass(intermediate2))))
> [1] "matrix" "list"  
> 
> 
> 
> 
>  
> ______________________________________________________________
> ______________________
> Be a PS3 game guru.
> 
> ______________________________________________
> R-help at 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.
> 
> 
> 

------------------------------------------------------------------------------
Notice:  This e-mail message, together with any attachments,...{{dropped}}