R range referencing with different dimensions not causing an error - r

Question

In using range referencing I normally expect to see an error or at least a warning message when the operations in '[' ']' do not match the dimensions of the parent object, however I have just discovered that I am not seeing said warnings and errors. Is there a setting for this or a way to force an error? Example:
x = 1:5
y = 10:12
x[y>10]
y[x>2]
likewise this applies to data frames and other R objects:
dat = data.frame(x=runif(100),y=1:100)
dat[sample(c(TRUE,FALSE),23),c(TRUE,FALSE)]
The silent repetition and truncation of the references to match the dimensions of the parent object is unexpected, having used R for years, I've somehow never noticed this before.
I'm using R Console (64-bit) 3.0.1 for Windows (could be updated yes, but I hope this isn't the cause).
Edit: Fixed data.frame example as data.frame's don't allow more column references than columns. Thanks zero323.

Answer 1

You could modify the `[.data.frame` function to throw a warning when indexing with a logical vector that doesn't evenly divide the number of rows:
`[.data.frame` <- function(x, i, j, drop = if (missing(i)) TRUE else length(cols) == 1) {
if (!missing(i) && is.logical(i) && nrow(x) %% length(i) != 0) {
warning("Indexing data frame with logical vector that doesn't evenly divide row count")
}
base::`[.data.frame`(x, i, j, drop)
}
Here's a demonstration with the 150-row iris dataset, passing logical indexing vectors of length 11 (should cause warning) and 15 (should not cause warning):
iris[c(rep(FALSE, 10), TRUE),]
# Sepal.Length Sepal.Width Petal.Length Petal.Width Species
# 11 5.4 3.7 1.5 0.2 setosa
# 22 5.1 3.7 1.5 0.4 setosa
# 33 5.2 4.1 1.5 0.1 setosa
# 44 5.0 3.5 1.6 0.6 setosa
# 55 6.5 2.8 4.6 1.5 versicolor
# 66 6.7 3.1 4.4 1.4 versicolor
# 77 6.8 2.8 4.8 1.4 versicolor
# 88 6.3 2.3 4.4 1.3 versicolor
# 99 5.1 2.5 3.0 1.1 versicolor
# 110 7.2 3.6 6.1 2.5 virginica
# 121 6.9 3.2 5.7 2.3 virginica
# 132 7.9 3.8 6.4 2.0 virginica
# 143 5.8 2.7 5.1 1.9 virginica
# Warning message:
# In `[.data.frame`(iris, c(rep(FALSE, 10), TRUE), ) :
# Indexing data frame with logical vector that doesn't evenly divide number of rows
iris[c(rep(FALSE, 14), TRUE),]
# Sepal.Length Sepal.Width Petal.Length Petal.Width Species
# 15 5.8 4.0 1.2 0.2 setosa
# 30 4.7 3.2 1.6 0.2 setosa
# 45 5.1 3.8 1.9 0.4 setosa
# 60 5.2 2.7 3.9 1.4 versicolor
# 75 6.4 2.9 4.3 1.3 versicolor
# 90 5.5 2.5 4.0 1.3 versicolor
# 105 6.5 3.0 5.8 2.2 virginica
# 120 6.0 2.2 5.0 1.5 virginica
# 135 6.1 2.6 5.6 1.4 virginica
# 150 5.9 3.0 5.1 1.8 virginica

Answer 2

Expanding on #josilber I've written the following for atomic vector and matrix subsetting in case anyone else wants it:
`[` <- function(x, i) {
if(!missin
g(i) && is.logical(i) && (length(x) %% length(i) != 0 || length(i) > length(x))) {
warning("Indexing atomic vector with logical vector that doesn't evenly divide row count")
}
base::`[`(x,i)
}
`[` <- function(x,i,j,...,drop=TRUE) {
if (!missing(i) && is.logical(i) && nrow(x) %% length(i) != 0) {
warning("Indexing matrix with logical vector that doesn't evenly divide row count")
}
if (!missing(j) && is.logical(j) && nrow(x) %% length(j) != 0) {
warning("Indexing matrix with logical vector that doesn't evenly divide column count")
}
base::`[`(x,i,j,...,drop)
}
Testing my original example afterwards with this modification now produces the warning and other operations behave as per normal:
> x =
1:5
> y = 10:12
> x[y>10]
[1] 2 3 5
Warning message:
In x[y > 10] :
Indexing atomic vector with logical vector that doesn't evenly divide row count
> y[x>2]
[1] 12 NA NA
Warning message:
In y[x > 2] :
Indexing atomic vector with logical vector that doesn't evenly divide row count
> x[x>2]
[1] 3 4 5
> x[1:2]
[1] 1 2