ROBUSTNESS: x || y and x && y to give warning/error if length(x) != 1 or length(y) != 1

[HenrikBengtsson]

Idea

In the spirit of Issue #38 (if/while (c(TRUE, TRUE)) ...) of giving a warning (soon error), @hadley proposed in a Tweet:

@HenrikBengtsson as part of new if() warning, I wonder if && and || should give warning when collapsing vector to scalar

Issue

Today we have that x || y performs x[1] || y for length(x) > 1. For instance,

> c(TRUE, TRUE) || FALSE
[1] TRUE
> c(TRUE, FALSE) || FALSE
[1] TRUE
> c(TRUE, NA) || FALSE
[1] TRUE
> c(FALSE, TRUE) || FALSE
[1] FALSE

This property is symmetric in LHS and RHS (i.e. y || x behaves the same) and it also applies to x && y.

The issue is that the above truncation of x is completely silent -there's neither an error nor a warning being produced.

Discussion/Suggestion

Using x || y and x && y with a non-scalar x or y is likely a mistake. Either the code is written assuming x and y are scalars, or there is a coding error and vectorized versions x | y and x & y were intended. Should x || y always be considered an mistake if length(x) != 1 or length(y) != 1? If so, should it be a warning or an error? For instance,

> x <- c(TRUE, TRUE)
> y <- FALSE
> x || y

Error in x || y : applying scalar operator || to non-scalar elements
Execution halted

What about the case where length(x) == 0 or length(y) == 0? Today x || y returns NA in such cases, e.g.

> logical(0) || c(FALSE, NA)
[1] NA
> logical(0) || logical(0)
[1] NA
> logical(0) && logical(0)
[1] NA

I don't know the background for this behavior, but I'm sure there is an argument behind that one. Maybe it's simply that || and && should always return a scalar logical and neither TRUE nor FALSE can be returned.

[brodieG]

I would suggest a warning for symmetry with:

> 1:2 + 1:3
[1] 2 4 4
Warning message:
In 1:2 + 1:3 :
  longer object length is not a multiple of shorter object length

Seems like a similar type of error. Changing the NA behavior seems like a much higher hurdle because if anyone ran into this in the past they would have noticed (e.g. if(NA)) so existing code may actually expet the NA in some cases.

I'm happy to help with this to the extent my abilities and other demands allow me to.

[HenrikBengtsson]

UPDATE: I just posted 'ROBUSTNESS: x || y and x && y to give warning/error if length(x) != 1 or length(y) != 1' to R-devel on 2018-08-28.

[HenrikBengtsson]

UPDATE 2018-09-12: R-devel r75289 just gained support for this:

NEWS: Experimentally, setting environment variable _R_CHECK_LENGTH_1_LOGIC2_ will lead to warnings (or errors if the variable is set to a 'true' value) when && or || encounter and use arguments of length more than one.

[HenrikBengtsson]

Hmm...

The _R_CHECK_LENGTH_1_LOGIC2_=true check on LHS || RHS applies only to the LHS and not the RHS:

> Sys.setenv("_R_CHECK_LENGTH_1_LOGIC2_"="true")
> c(TRUE, TRUE) || TRUE
Error in c(TRUE, TRUE) || TRUE : 
  'length(x) = 2 > 1' in coercion to 'logical(1)'

> TRUE || c(TRUE, TRUE)
[1] TRUE

For LHS && RHS it works as expected, e.g.

> c(TRUE, TRUE) && TRUE
Error in c(TRUE, TRUE) && TRUE : 
  'length(x) = 2 > 1' in coercion to 'logical(1)'

> TRUE && c(TRUE, TRUE)
Error in TRUE && c(TRUE, TRUE) : 
  'length(x) = 2 > 1' in coercion to 'logical(1)'

Verified on R version 3.6.1 (2019-07-05), R version 3.6.1 Patched (2019-09-12 r77183), and R Under development (unstable) (2019-10-12 r77279).

UPDATE 2019-10-12: I've filed PR17630 about this problem.

UPDATE 2019-10-12: This comment was invalid because:

Doh... brain fried. To further clarify Peter's point/explanation for future eyes. The following is indeed expected:

Sys.setenv("R_CHECK_LENGTH_1_LOGIC2"="true")
TRUE || c(TRUE, TRUE)
[1] TRUE
FALSE || c(TRUE, TRUE)
Error in FALSE || c(TRUE, TRUE) :
'length(x) = 2 > 1' in coercion to 'logical(1)'

and

FALSE && c(TRUE, TRUE)
[1] FALSE
TRUE && c(TRUE, TRUE)
Error in TRUE && c(TRUE, TRUE) :
'length(x) = 2 > 1' in coercion to 'logical(1)'

It's only when R needs evaluates the RHS, to resolve 'LHS || RHS' or 'LHS && RHS', that it will be evaluated(*). Without evaluating RHS, it is not possible to know length(RHS). (I'm pretty sure this was also discussed in length before on R-devel when 'R_CHECK_LENGTH_1_LOGIC2' was first proposed.)

(*) This is document in help("Logic", package="base") as:

The longer form [&& and ||] evaluates left to right examining only the first
element of each vector. Evaluation proceeds only until the result is
determined.

[HenrikBengtsson]

UPDATE 2022-02-23: _R_CHECK_LENGTH_1_LOGIC2_=warn is the new default in R-devel (to become R 4.2.0), cf. wch/r-source@b3b9f66

[HenrikBengtsson]

UPDATE 2022-04-27: This will now be an error in R-devel (to become R 4.3.0), cf. Make calling && or || with either argument of length greater than one into an error.