R help - Aug 2011 - Question about object permanence/marshalling

Hello,
 
I am trying to write some code that dumps R objects to the harddisk in a binary
format so they can be quickly re-used later. Goal is to save time. The objects
may be quite large (e.g. classes for a GUI). I was thinking that save() and
load() would be suitable for this (until now I only thought it could be used for
'real' data, e.g. matrices, data.frames etc), but I am hoping any object
can be 'marshalled' using these functions. Probably I am doing something
wrong in the unmarshal() function, perhaps with assign().
 
Thank you in advance!
 
AJ
 
 
#########
# Creation of test data
#########
setClass(
  Class="Test",
    representation=representation(
      amounts="data.frame"
      )
)
setMethod(
  f="initialize",
  signature="Test",
  definition=function(.Object, amounts){
    .Object@amounts <- amounts
    return(.Object)
  }
)
setGeneric (
  name="doStuff",
  def=function(.Object){standardGeneric("doStuff")}
)
setMethod(
  f = "doStuff",
  signature = "Test",
  definition=function(.Object) {
    return(mean(.Object@amounts, na.rm=TRUE))
  }
)
    
print( objects() )
instance <- new(Class="Test", data.frame(amount=runif(10, 0, 10)))
doStuff(instance) 
 
#########
# actual code (incomplete)
#########
marshal <- function(object) {
    fn <- file.path(Sys.getenv()["TEMP"], paste(object,
".xdr", sep=""))
    save(object, file=fn, compress=FALSE)
    print(sprintf("Saving %s", fn))
    }
unmarshal <- function(xdr) {
    object <- strsplit(strsplit(xdr, "\\.")[[1]][[1]],
"/")
    object <- object[[1]][length(object[[1]])]
    assign(object, load(xdr))
    print(sprintf("Loading %s", xdr))
    }
print(objects())
lapply(c("doStuff", "instance"), marshal)
rm(list=c("doStuff", "instance"))
   
xdrs <- Sys.glob(file.path(Sys.getenv()["TEMP"],
"*.xdr"))
lapply(xdrs, unmarshal)
print(objects())  ## doStuff and instance do not appear! :-(


Cheers!!
Albert-Jan


~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
All right, but apart from the sanitation, the medicine, education, wine, public
order, irrigation, roads, a fresh water system, and public health, what have the
Romans ever done for us?
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
	[[alternative HTML version deleted]]

The problem I think is in your unmarshal.  'load' will load the object
into the local environment, not the global.  You have to explicitly
return it, that means you have to know the name that its was 'save'd
by;
> unmarshal <- function(xdr) {
> ??? object <- strsplit(strsplit(xdr, "\\.")[[1]][[1]],
"/")
> ??? object <- object[[1]][length(object[[1]])]
> ??? load(xdr)
> ??? print(sprintf("Loading %s", xdr))
       ?????  # return the object that was just loaded; don't know if
they all have the same name.
> ??? }



On Thu, Aug 25, 2011 at 9:34 AM, Albert-Jan Roskam <fomcl at yahoo.com>
wrote:
> Hello,
>
> I am trying to write some code that dumps R objects to the harddisk in a
binary format so they can be quickly re-used later. Goal is to save time. The
objects may be quite large (e.g. classes for a GUI). I was thinking that save()
and load() would be suitable for this (until now I only thought it could be used
for 'real' data, e.g. matrices, data.frames etc), but I am hoping any
object can be 'marshalled' using these functions. Probably I am doing
something wrong in the unmarshal() function, perhaps with assign().
>
> Thank you in advance!
>
> AJ
>
>
> #########
> # Creation of test data
> #########
> setClass(
> ? Class="Test",
> ??? representation=representation(
> ????? amounts="data.frame"
> ????? )
> )
> setMethod(
> ? f="initialize",
> ? signature="Test",
> ? definition=function(.Object, amounts){
> ??? .Object at amounts <- amounts
> ??? return(.Object)
> ? }
> )
> setGeneric (
> ? name="doStuff",
> ? def=function(.Object){standardGeneric("doStuff")}
> )
> setMethod(
> ? f = "doStuff",
> ? signature = "Test",
> ? definition=function(.Object) {
> ??? return(mean(.Object at amounts, na.rm=TRUE))
> ? }
> )
>
> print( objects() )
> instance <- new(Class="Test", data.frame(amount=runif(10, 0,
10)))
> doStuff(instance)
>
> #########
> # actual code (incomplete)
> #########
> marshal <- function(object) {
> ??? fn <- file.path(Sys.getenv()["TEMP"], paste(object,
".xdr", sep=""))
> ??? save(object, file=fn, compress=FALSE)
> ??? print(sprintf("Saving %s", fn))
> ??? }
> unmarshal <- function(xdr) {
> ??? object <- strsplit(strsplit(xdr, "\\.")[[1]][[1]],
"/")
> ??? object <- object[[1]][length(object[[1]])]
> ??? assign(object, load(xdr))
> ??? print(sprintf("Loading %s", xdr))
> ??? }
> print(objects())
> lapply(c("doStuff", "instance"), marshal)
> rm(list=c("doStuff", "instance"))
>
> xdrs <- Sys.glob(file.path(Sys.getenv()["TEMP"],
"*.xdr"))
> lapply(xdrs, unmarshal)
> print(objects())? ## doStuff and instance do not appear! :-(
>
>
> Cheers!!
> Albert-Jan
>
>
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> All right, but apart from the sanitation, the medicine, education, wine,
public order, irrigation, roads, a fresh water system, and public health, what
have the Romans ever done for us?
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> ? ? ? ?[[alternative HTML version deleted]]
>
>
> ______________________________________________
> R-help at r-project.org 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.
>
>


-- 
Jim Holtman
Data Munger Guru

What is the problem that you are trying to solve?