[Rd] Recent changes to as.complex(NA_real_)

Spencer Graves @pencer@gr@ve@ @end|ng |rom prod@y@e@com
Fri Sep 22 23:42:03 CEST 2023 

	  Perhaps I shouldn't comment without having read the entire thread, 
but I will:  I can envision situations where I might want, e.g., 2 from 
complex(r=2, i=NA_real_).


	  Spencer Graves


On 9/22/23 3:43 PM, Duncan Murdoch wrote:
> Since the result of is.na(x) is the same on each of those, I don't see a 
> problem.  As long as that is consistent, I don't see a problem.  You 
> shouldn't be using any other test for NA-ness.  You should never be 
> expecting identical() to treat different types as the same (e.g. 
> identical(NA, NA_real_) is FALSE, as it should be).  If you are using a 
> different test, that's user error.
> 
> Duncan Murdoch
> 
> On 22/09/2023 2:41 p.m., Hervé Pagès wrote:
>> We could also question the value of having an infinite number of NA
>> representations in the complex space. For example all these complex
>> values are displayed the same way (as NA), are considered NAs by
>> is.na(), but are not identical or semantically equivalent (from an Re()
>> or Im() point of view):
>>
>>       NA_real_ + 0i
>>
>>       complex(r=NA_real_, i=Inf)
>>
>>       complex(r=2, i=NA_real_)
>>
>>       complex(r=NaN, i=NA_real_)
>>
>> In other words, using a single representation for complex NA (i.e.
>> complex(r=NA_real_, i=NA_real_)) would avoid a lot of unnecessary
>> complications and surprises.
>>
>> Once you do that, whether as.complex(NA_real_) should return
>> complex(r=NA_real_, i=0) or complex(r=NA_real_, i=NA_real_) becomes a
>> moot point.
>>
>> Best,
>>
>> H.
>>
>> On 9/22/23 03:38, Martin Maechler wrote:
>>>>>>>> Mikael Jagan
>>>>>>>>       on Thu, 21 Sep 2023 00:47:39 -0400 writes:
>>>       > Revisiting this thread from April:
>>>
>>>       >https://stat.ethz.ch/pipermail/r-devel/2023-April/082545.html
>>>
>>>       > where the decision (not yet backported) was made for
>>>       > as.complex(NA_real_) to give NA_complex_ instead of
>>>       > complex(r=NA_real_, i=0), to be consistent with
>>>       > help("as.complex") and as.complex(NA) and 
>>> as.complex(NA_integer_).
>>>
>>>       > Was any consideration given to the alternative?
>>>       > That is, to changing as.complex(NA) and 
>>> as.complex(NA_integer_) to
>>>       > give complex(r=NA_real_, i=0), consistent with
>>>       > as.complex(NA_real_), then amending help("as.complex")
>>>       > accordingly?
>>>
>>> Hmm, as, from R-core, mostly I was involved, I admit to say "no",
>>> to my knowledge the (above) alternative wasn't considered.
>>>
>>>     > The principle that
>>>     > Im(as.complex(<real=(double|integer|logical)>)) should be zero
>>>     > is quite fundamental, in my view, hence the "new" behaviour
>>>     > seems to really violate the principle of least surprise ...
>>>
>>> of course "least surprise"  is somewhat subjective.  Still,
>>> I clearly agree that the above would be one desirable property.
>>>
>>> I think that any solution will lead to *some* surprise for some
>>> cases, I think primarily because there are *many* different
>>> values z  for which  is.na(z)  is true,  and in any case
>>> NA_complex_  is only of the many.
>>>
>>> I also agree with Mikael that we should reconsider the issue
>>> that was raised by Davis Vaughan here ("on R-devel") last April.
>>>
>>>       > Another (but maybe weaker) argument is that
>>>       > double->complex coercions happen more often than
>>>       > logical->complex and integer->complex ones.  Changing the
>>>       > behaviour of the more frequently performed coercion is
>>>       > more likely to affect code "out there".
>>>
>>>       > Yet another argument is that one expects
>>>
>>>       >      identical(as.complex(NA_real_), NA_real_ + (0+0i))
>>>
>>>       > to be TRUE, i.e., that coercing from double to complex is
>>>       > equivalent to adding a complex zero.  The new behaviour
>>>       > makes the above FALSE, since NA_real_ + (0+0i) gives
>>>       > complex(r=NA_real_, i=0).
>>>
>>> No!  --- To my own surprise (!) --- in current R-devel the above is 
>>> TRUE,
>>> and
>>>         NA_real_ + (0+0i)  , the same as
>>>         NA_real_ + 0i      , really gives  complex(r=NA, i=NA) :
>>>
>>> Using showC() from ?complex
>>>
>>>     showC <- function(z) noquote(sprintf("(R = %g, I = %g)", Re(z), 
>>> Im(z)))
>>>
>>> we see (in R-devel) quite consistently
>>>
>>>> showC(NA_real_ + 0i)
>>> [1] (R = NA, I = NA)
>>>> showC(NA       + 0i)  # NA is 'logical'
>>> [1] (R = NA, I = NA)
>>> where as in R 4.3.1 and "R-patched" -- *in*consistently
>>>
>>>> showC(NA_real_ + 0i)
>>> [1] (R = NA, I = 0)
>>>> showC(NA + 0i)
>>> [1] (R = NA, I = NA)
>>> .... and honestly, I do not see *where* (and when) we changed
>>> the underlying code (in arithmetic.c !?)  in R-devel to *also*
>>> produce  NA_complex_  in such complex *arithmetic*
>>>
>>>
>>>       > Having said that, one might also (but more naively) expect
>>>
>>>       >     identical(as.complex(as.double(NA_complex_)), NA_complex_)
>>>
>>>       > to be TRUE.
>>>
>>> as in current R-devel
>>>
>>>       > Under my proposal it continues to be FALSE.
>>>
>>> as in "R-release"
>>>
>>>       > Well, I'd prefer if it gave FALSE with a warning
>>>       > "imaginary parts discarded in coercion", but it seems that
>>>       > as.double(complex(r=a, i=b)) never warns when either of
>>>       > 'a' and 'b' is NA_real_ or NaN, even where "information"
>>>       > {nonzero 'b'} is clearly lost ...
>>>
>>> The question of *warning* here is related indeed, but I think
>>> we should try to look at it only *secondary* to your first
>>> proposal.
>>>
>>>       > Whatever decision is made about as.complex(NA_real_),
>>>       > maybe these points should be weighed before it becomes part of
>>>       > R-release ...
>>>
>>>       > Mikael
>>>
>>> Indeed.
>>>
>>> Can we please get other opinions / ideas here?
>>>
>>> Thank you in advance for your thoughts!
>>> Martin
>>>
>>> ---
>>>
>>> PS:
>>>
>>>    Our *print()*ing  of complex NA's ("NA" here meaning NA or NaN)
>>>    is also unsatisfactory, e.g. in the case where all entries of a
>>>    vector are NA in the sense of is.na(.), but their
>>>    Re() and Im() are not all NA:
>>>     showC <- function(z) noquote(sprintf("(R = %g, I = %g)", Re(z), 
>>> Im(z)))
>>>     z <- complex(, c(11, NA, NA), c(NA, 99, NA))
>>>     z
>>>     showC(z)
>>>
>>> gives
>>>
>>>     > z
>>>     [1] NA NA NA
>>>     > showC(z)
>>>     [1] (R = 11, I = NA) (R = NA, I = 99) (R = NA, I = NA)
>>>
>>> but that (printing of complex) *is* another issue,
>>> in which we have the re-opened bugzilla PR#16752
>>>       ==>https://bugs.r-project.org/show_bug.cgi?id=16752
>>>
>>> on which we also worked during the R Sprint in Warwick three
>>> weeks ago, and where I want to commit changes in any case {but
>>> think we should change even a bit more than we got to during the
>>> Sprint}.
>>>
>>> ______________________________________________
>>> R-devel using r-project.org  mailing list
>>> https://stat.ethz.ch/mailman/listinfo/r-devel
>>
> 
> ______________________________________________
> R-devel using r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-devel

More information about the R-devel mailing list