R help - Jan 2004 - R internal data types

I am trying to figure out R data types and/or storage mode.  For example:
> #From a clean workspace
> gc()
         used (Mb) gc trigger (Mb)
Ncells 415227 11.1     597831   16
Vcells 103533  0.8     786432    6
> x <- seq(0,100000,1)
> is.integer(x)
[1] FALSE
> is.double(x)
[1] TRUE
> object.size(x)
[1] 800036
> gc()
         used (Mb) gc trigger (Mb)
Ncells 415247 11.1     667722 17.9
Vcells 203543  1.6     786432  6.0
> x <- as.integer(x)
> is.integer(x)
[1] TRUE
> is.double(x)
[1] FALSE
> gc()
         used (Mb) gc trigger (Mb)
Ncells 415249 11.1     741108 19.8
Vcells 153543  1.2     786432  6.0
> x <- 1:100000
> is.integer(x)
> gc()
         used (Mb) gc trigger (Mb)
Ncells 415278 11.1     741108 19.8
Vcells 153553  1.2     786432  6.0
> is.integer(3)
[1] FALSE
> is.double(3)
[1] TRUE
> is.integer(3 * as.integer(5))
[1] FALSE
> is.integer(as.integer(3) * as.integer(5))
[1] TRUE
> is.integer(c(as.integer(5),as.integer(6),as.integer(7)))
[1] TRUE
> is.integer(c(as.integer(5),as.integer(6),7))
[1] FALSE
> is.integer(seq(as.integer(5),as.integer(10),1))
[1] FALSE
> is.integer(seq(as.integer(5),as.integer(10),as.integer(1)))
[1] TRUE

So it looks like R stores numbers as doubles unless the are converted to
integers (long) with the as.integer() function or they are created with the
: operator.  If any of the numbers to a function are not type integer than
the function returns type double.  Is this the case?  Thanks.

Ben Stabler
Oregon Department of Transportation

>>>>> "Benjamin" == Benjamin STABLER
<Benjamin.STABLER at odot.state.or.us>
>>>>>     on Wed, 14 Jan 2004 15:52:46 -0800 writes:
	  <........>

    Benjamin> So it looks like R stores numbers as doubles
    Benjamin> unless the are converted to integers (long) with
    Benjamin> the as.integer() function or they are created with
    Benjamin> the : operator.  
+/- yes;
In most cases, this should not matter though.

There are a few other functions that return integer ``by
definition'', e.g.,
length(), dim(), nrow(), ncol()  {the latter 3 return 'NULL' or an
integer}.


    Benjamin> If any of the numbers to a function are not type
    Benjamin> integer than the function returns type double.  Is
    Benjamin> this the case?

Most functions will return double even when all numeric arguments are
integer.  Also, e.g. length() will return integer() in any case.

Only for some "arithmetic" functions
      (+, -, *, %%, %/%, also min(), max(), sum())
your statement is true.
But really, in most cases, code shouldn't rely on this.

Martin Maechler <maechler at stat.math.ethz.ch>
http://stat.ethz.ch/~maechler/
Seminar fuer Statistik, ETH-Zentrum  LEO C16	Leonhardstr. 27
ETH (Federal Inst. Technology)	8092 Zurich	SWITZERLAND
phone: x-41-1-632-3408		fax: ...-1228			<><

On Wed, 14 Jan 2004 Benjamin.STABLER at odot.state.or.us wrote:
> I am trying to figure out R data types and/or storage mode.  For example:
[...]
> So it looks like R stores numbers as doubles unless the are converted to
> integers (long) with the as.integer() function or they are created with the
> : operator.  If any of the numbers to a function are not type integer than
> the function returns type double.  Is this the case?  Thanks.
Nearly!

Non-complex `numbers' can have storage mode "double" or
"integer".
Storage mode "integer" is the C `int' type and not the C
`long' type, so
probably on all current R platforms doubles are stored in 8 bytes and
integers in 4.

The storage mode is largely under the user/programmer's control: you can 
do

storage.mode(x) <- "double"

for example.  The storage mode of the return value of a function depends 
on how it was programmed, not on the types of its arguments, for example
> x <- seq(10)
> storage.mode(x)
[1] "integer"

Just a few base functions normally return integer results: you have 
mentioned : and as.integer() and I have illustrated seq().  table() and 
tabulate() are two others I know of, and factors are integer vectors with 
particular attributes.

Current versions of S use integer storage mode much more widely, and it is 
quite possible that in due course more functions in R will make use of it.

-- 
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