[[ vs. .subset2

[Gabor Grothendieck]

In the code below test2() gives an error message:

   Error in .subset2(x, i, exact = exact) :
     attempt to select less than one element

even though test(), which is nearly the same, gives
the expected result.  BOD is a data set that comes
with R.   Is this a bug?


idx <- 2

# returns expected result
test <- function(pf = parent.frame()) .subset2(BOD, pf$idx)
test()

# gives error message !!!
test2 <- function(pf = parent.frame()) BOD[[pf$idx]]
test2()


I tried this on Windows Vista under:

> R.version.string
[1] "R version 2.7.1 RC (2008-06-16 r45949)"

and

> R.version.string
[1] "R version 2.8.0 Under development (unstable) (2008-06-28 r46012)"

______________________________________________
R-devel@... mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel

[Spencer Graves]

Dear Gabor:

      Does the following answer your question:

 > tst <- list(a=1, b=2)
 > tst[numeric(0)]
list()
 > tst[[numeric(0)]]
Error in tst[[numeric(0)]] : attempt to select less than one element

      Spencer

Gabor Grothendieck wrote:

> In the code below test2() gives an error message:
>
>    Error in .subset2(x, i, exact = exact) :
>      attempt to select less than one element
>
> even though test(), which is nearly the same, gives
> the expected result.  BOD is a data set that comes
> with R.   Is this a bug?
>
>
> idx <- 2
>
> # returns expected result
> test <- function(pf = parent.frame()) .subset2(BOD, pf$idx)
> test()
>
> # gives error message !!!
> test2 <- function(pf = parent.frame()) BOD[[pf$idx]]
> test2()
>
>
> I tried this on Windows Vista under:
>
>  
>> R.version.string
>>    
> [1] "R version 2.7.1 RC (2008-06-16 r45949)"
>
> and
>
>  
>> R.version.string
>>    
> [1] "R version 2.8.0 Under development (unstable) (2008-06-28 r46012)"
>
> ______________________________________________
> R-devel@... mailing list
> https://stat.ethz.ch/mailman/listinfo/r-devel
>

______________________________________________
R-devel@... mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel

[Gabor Grothendieck]

Not really.  The point of this is that promises seem to be
screwing up in conjunction with [[ but in light of the fact
that .subset2 works it suggests that there might be a better
implementation of [[.

Also some insight into what is really driving this problem
so one can more easily recognize it in other places would
be helpful.

This a simplification of an actual problem I had in
a large program and it took some time to track this
one down.  Also, I am not entirely happy with the
workarounds which are to use .subset2, as shown,
or force(pf) since it seems more natural to me that
it should just work out of the box.

On Fri, Jul 4, 2008 at 8:42 PM, Spencer Graves <spencer.graves@...> wrote:

> Dear Gabor:
>     Does the following answer your question:
>> tst <- list(a=1, b=2)
>> tst[numeric(0)]
> list()
>> tst[[numeric(0)]]
> Error in tst[[numeric(0)]] : attempt to select less than one element
>
>     Spencer
>
> Gabor Grothendieck wrote:
>>
>> In the code below test2() gives an error message:
>>
>>   Error in .subset2(x, i, exact = exact) :
>>     attempt to select less than one element
>>
>> even though test(), which is nearly the same, gives
>> the expected result.  BOD is a data set that comes
>> with R.   Is this a bug?
>>
>>
>> idx <- 2
>>
>> # returns expected result
>> test <- function(pf = parent.frame()) .subset2(BOD, pf$idx)
>> test()
>>
>> # gives error message !!!
>> test2 <- function(pf = parent.frame()) BOD[[pf$idx]]
>> test2()
>>
>>
>> I tried this on Windows Vista under:
>>
>>
>>>
>>> R.version.string
>>>
>>
>> [1] "R version 2.7.1 RC (2008-06-16 r45949)"
>>
>> and
>>
>>
>>>
>>> R.version.string
>>>
>>
>> [1] "R version 2.8.0 Under development (unstable) (2008-06-28 r46012)"
>>
>> ______________________________________________
>> R-devel@... mailing list
>> https://stat.ethz.ch/mailman/listinfo/r-devel
>>
>

______________________________________________
R-devel@... mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel

[Peter Dalgaard]

Gabor Grothendieck wrote:
> Not really.  The point of this is that promises seem to be
> screwing up in conjunction with [[ but in light of the fact
> that .subset2 works it suggests that there might be a better
> implementation of [[
Not really; .subset2() is a shortcut that bypasses method dispatch, so
not a drop-in replacement for "[[". You also have

 > test3 <- function(pf = parent.frame()) "[[.data.frame"(BOD, pf$idx)
 > test3()
[1]  8.3 10.3 19.0 16.0 15.6 19.8


Something appears not quite right, but an easy workaround is to force
the promise:

 > test4 <- function(pf = parent.frame()) {pf;BOD[[pf$idx]]}
 > test4()
[1]  8.3 10.3 19.0 16.0 15.6 19.8



--
   O__  ---- Peter Dalgaard             Øster Farimagsgade 5, Entr.B
  c/ /'_ --- Dept. of Biostatistics     PO Box 2099, 1014 Cph. K
 (*) \(*) -- University of Copenhagen   Denmark      Ph:  ([phone]
~~~~~~~~~~ - (p.dalgaard@...)              FAX: [phone]

______________________________________________
R-devel@... mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel