R help - Nov 2006 - reaccessing array at a later date - trying to write it to file

Dear R-help community

I am trying to write an R object (data.out) to a file in order to re-access it 
later and not have to re-load up the array with data every time. Here is the 
form of data.out   
> data.out <- list(lats=seq(88.542, -88.542, length=94),
		  lons=seq(0, 360-1.875, length=192),
		  date=vector(length=nyr*12),
		  data=array(NA, c(nyr*12, 94*192))
) 
		  
I tried 
> save(data.out, file="/home/jenny/data/data.out.RData",
ascii=TRUE) and
combination of ways to re-access it but I couldn't reaccess it and therefore
use
the data within. 

I then tried 
>
dput(data.out,file="/home/jmb/sst_precip/gribdata/data.out.RData",
control="showAttributes")
#but the associated dget seemed to freeze my terminal window up and took much 
too long (I gave up after about 15 minutes) to read this file.

Does anybody have an idea about another way (or a way within these methods) to 
come back to my data within the list data.out (with all of its associated 
data.out$lats, data.out$lons, data.out$date, data.out$data) at a later stage to 
use the data within it to actually get on with some analysis? 

I appreciate your time in reading this and look forward to hearing any 
suggestions,

Sincerely,

Jenny Barnes






~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Jennifer Barnes
PhD student - long range drought prediction
Climate Extremes
Department of Space and Climate Physics
University College London
Holmbury St Mary, Dorking
Surrey
RH5 6NT
01483 204149
07916 139187
Web: http://climate.mssl.ucl.ac.uk

On Thu, 23 Nov 2006, Jenny Barnes wrote:
> Dear R-help community
>
> I am trying to write an R object (data.out) to a file in order to re-access
it
> later and not have to re-load up the array with data every time. Here is
the
> form of data.out
>
>> data.out <- list(lats=seq(88.542, -88.542, length=94),
> 		  lons=seq(0, 360-1.875, length=192),
> 		  date=vector(length=nyr*12),
> 		  data=array(NA, c(nyr*12, 94*192))
> )
>
> I tried
>> save(data.out, file="/home/jenny/data/data.out.RData",
ascii=TRUE) and
> combination of ways to re-access it but I couldn't reaccess it and
therefore use
> the data within.
What stopped you re-accessing it?  I would use

save(data.out, file="/home/jenny/data/data.out.RData")
....
load("/home/jenny/data/data.out.RData")

If that does not work, we need to see the transcript to (perhaps) 
understand why (and all the usual details about your environment: it is 
possible to save really large objects that you cannot restore on a 
32-bit machine).

-- 
Brian D. Ripley,                  ripley at stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595

Thank you for your response Professor Ripley,

Having tried again your suggestion of load() worked (well - it finished, which I
assume it meant it worked). However not I am confused as to how I can check it 
has worked. 
I typed
>data.out$data
which called up the data from the file - but I'm not sure if this is data
from
the file I have just restored as in my "previously saved workspace
restored"
data.out is still there so if I typed data.out$data I wouldn't know if it
was
coming from the newly loaded file or from previously. Is there an alternative 
way to check it has loaded properly? Also, is it normal that if I type
>data.out.RData
it says
Error: object "data.out.RData" not found

Here are my details:

platform       sparc-sun-solaris2.10     
arch           sparc                     
os             solaris2.10               
system         sparc, solaris2.10        
status                                   
major          2                         
minor          3.1                       
year           2006                      
month          06                        
day            01                        
svn rev        38247                     
language       R                         
version.string Version 2.3.1 (2006-06-01)

My computer shouldn't have a problem with dealing with this data.
>On Thu, 23 Nov 2006, Jenny Barnes wrote:
>
>> Dear R-help community
>>
>> I am trying to write an R object (data.out) to a file in order to
re-access
it
>> later and not have to re-load up the array with data every time. Here
is the
>> form of data.out
>>
>>> data.out <- list(lats=seq(88.542, -88.542, length=94),
>> 		  lons=seq(0, 360-1.875, length=192),
>> 		  date=vector(length=nyr*12),
>> 		  data=array(NA, c(nyr*12, 94*192))
>> )
>>
>> I tried
>>> save(data.out, file="/home/jenny/data/data.out.RData",
ascii=TRUE) and
>> combination of ways to re-access it but I couldn't reaccess it and
therefore
use
>> the data within.
>
>What stopped you re-accessing it?  I would use
>
>save(data.out, file="/home/jenny/data/data.out.RData")
>....
>load("/home/jenny/data/data.out.RData")
>
>If that does not work, we need to see the transcript to (perhaps) 
>understand why (and all the usual details about your environment: it is 
>possible to save really large objects that you cannot restore on a 
>32-bit machine).
>
>-- 
>Brian D. Ripley,                  ripley at stats.ox.ac.uk
>Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
>University of Oxford,             Tel:  +44 1865 272861 (self)
>1 South Parks Road,                     +44 1865 272866 (PA)
>Oxford OX1 3TG, UK                Fax:  +44 1865 272595
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Jennifer Barnes
PhD student - long range drought prediction
Climate Extremes
Department of Space and Climate Physics
University College London
Holmbury St Mary, Dorking
Surrey
RH5 6NT
01483 204149
07916 139187
Web: http://climate.mssl.ucl.ac.uk

Thank you Barry for your time in responding!

I think that will really help - the difference between attach and load were not 
clear to me before your reply! Also I did not know about rm() - thank you for 
the detail, I know you took longer than you had planned but I do appreciate it,

For those with a similar problems in the future please see the responses below:
>
>Jenny Barnes wrote:
>
>> Having tried again your suggestion of load() worked (well - it
finished,
which I 
>> assume it meant it worked). However not I am confused as to how I can
check
it 
>> has worked. 
>> I typed
>> 
>>>data.out$data
>> 
>> which called up the data from the file - but I'm not sure if this
is data
from 
>> the file I have just restored as in my "previously saved workspace
restored"
>
>  Remove it from your current workspace:
>
>  > rm(data.out)
>
>  then do the load('whatever') again:
>
>  > load("/some/path/to/data.out.RData")
>
>  then see if its magically re-appeared in your workspace:
>
>  > data.out$data
>
>  But now if you quit and save your workspace it'll be in your workspace
>again when you start up.
>
>  So you could consider 'attach' instead of 'load'...
>
>  Remove data.out from your current workspace, save your current 
>workspace (with 'save()' - just like that with nothing in the 
>parentheses), then instead of load('/some/path/to/data.out.RData')
use:
>
>  > attach('/some/path/to/data.out.RData')
>
>  This makes R search for an object called 'data.out' in that file 
>whenever you type 'data.out'. It will find it as long as there's
not a
>thing called 'data.out' in your workspace. So if you do attach(...)
and
>then do:
>
>  > str(data.out)
>
>  you'll see info about your data.out object, but then do:
>
>  > data.out=99
>  > str(data.out)
>
>  you'll see info about '99'. Your data.out is still happily
sitting in
>its .RData file, its just masked by the data.out we created and set to 
>99. Delete that, and your data.out comes back:
>
>  > rm(data.out)
>  > str(data.out) # - your data object again
>
>  The advantage of this is that data.out wont be stored in your current 
>workspace again. The disadvantage is that you have to do 
>'attach(...whatever...)' when you start R, and that data.out can be 
>masked if you create something with that name in your workspace. It is a 
>handy thing to do if you create large data objects that aren't going to 
>change much.
>
>>  Also, is it normal that if I type 
>> 
>>>data.out.RData
>> 
>> it says
>> Error: object "data.out.RData" not found
>
>  Yes, because thats the name of the _file_ on your computer and not the 
>R object.
>
>  This should be in the R manuals and help files... and I've gone on 
>much longer than I intended to in this email :)
>
>Barry
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Jennifer Barnes
PhD student - long range drought prediction
Climate Extremes
Department of Space and Climate Physics
University College London
Holmbury St Mary, Dorking
Surrey
RH5 6NT
01483 204149
07916 139187
Web: http://climate.mssl.ucl.ac.uk