How refer to variable name in R data frame if column does not have header?

Question

I'm using the nearZeroVar function in R to identify zero-variance variables in my analysis dataset.

An example of my code:

train1_nzv_y0 <- nearZeroVar(train1[train1$y=="no",], saveMetrics= TRUE)
train1_nzv_y0[train1_nzv_y0$zeroVar=="TRUE",]
           freqRatio percentUnique zeroVar  nzv
x1                 0  0.0001649599    TRUE TRUE
x2                 0  0.0001649599    TRUE TRUE
x3                 0  0.0001649599    TRUE TRUE
x4                 0  0.0001649599    TRUE TRUE

After producing the data frame train1_nzv_y0, I'd like to pull a list of the variable names x1-x4 which have been identified as zero-variance variables.

However, there is no column header for the variable names - they're "floating" in front of my first column (freqRatio) but there's no way to identify them.

Is there a standard trick in R for referring to ghost columns with missing headers?

Answer 1

They are the row names, so you can use rownames() to get those variable names:

rownames(train1_nzv_y0[train1_nzv_y0$zeroVar=="TRUE",])

This will return a character vector containing "x1", "x2", "x3", "x4".

Code

#install.packages("caret")
library(caret)

# Create example dataset
set.seed(123)
train1 <- data.frame(
  y = sample(c("yes", "no"), 100, replace=TRUE),
  x1 = c(rep(0, 50), rep(NA, 50)),  # zero variance for "no" class
  x2 = c(rep(0, 50), rep(1, 50)),   # normal variance
  x3 = c(rep(0, 50), rep(NA, 50)),  # zero variance for "no" class
  x4 = 1:100                         # normal variance
)

train1_nzv_y0 <- nearZeroVar(train1[train1$y=="no",], saveMetrics=TRUE)
train1_nzv_y0[train1_nzv_y0$zeroVar=="TRUE",]    
rownames(train1_nzv_y0[train1_nzv_y0$zeroVar=="TRUE",])

Output:

> train1_nzv_y0[train1_nzv_y0$zeroVar=="TRUE",]
   freqRatio percentUnique zeroVar  nzv
y          0      2.325581    TRUE TRUE
x1         0      2.325581    TRUE TRUE
x3         0      2.325581    TRUE TRUE
> 
> rownames(train1_nzv_y0[train1_nzv_y0$zeroVar=="TRUE",])
[1] "y"  "x1" "x3"