Check row-wise NA sum in R data.table

Question

Problem: I want to check if a row contains solely of NA's in a data.table object. Currently, I have an implementation depending on apply . Is there a more efficient while readable solution?

Any improvements and ideas are welcome! Thanks

dt <- data.table(
  x = c("A", "B", "C", "D"),
  y = c("true", NA, NA, "true"),
  z = c(NA, NA, "true", "true"),
  a = c(NA, NA, NA, "ha")
)

# Current Code:
idx <- apply(dt[, c(2:ncol(dt)), with = FALSE], 1, function(x) all(is.na(x)))
dt <- dt[!idx] 

# Code Attempt 1 (not so nice due to temp na_count column)
rel_cols <- names(dt)[!names(dt) %in% c("x")]
dt[, na_count := rowSums(is.na(.SD)), .SDcols = rel_cols][na_count < (ncol(dt) - 2)]

Answer 1

You can use rowSums like this :

library(data.table)
dt[rowSums(!is.na(dt[, ..rel_cols])) > 0]

#   x    y    z    a
#1: A true <NA> <NA>
#2: C <NA> true <NA>
#3: D true true   ha

Or using .SDcols :

dt[dt[, rowSums(!is.na(.SD)) > 0, .SDcols = rel_cols]]

Answer 2

Here is one base R option:

library(data.table)

dt[, rowSums(is.na(dt)) == ncol(dt)]

      x    y    z    a
1: <NA> <NA> <NA> <NA>

Data:

dt <- data.table(
    x = c("A", NA, "C", "D"),
    y = c("true", NA, NA, "true"),
    z = c(NA, NA, "true", "true"),
    a = c(NA, NA, NA, "ha")
)

Note: I intentionally slightly altered your sample data to make the second row of the data table all NA values, to demonstrate the answer is working.