How to get the number of columns with ONLY NA values?

Question

How to get the number of columns with ALL NA values? Meaning that every value in a column needs to be NA, for the column to be counted.

Answer 1

You could use colSums to count NA in the column and check if is the same as all row numbers of the dataframe. Here is a reproducible example where there are 2 columns with all values NA:

sum(colSums(is.na(df)) == nrow(df))
#> [1] 2

^{Created on 2023-01-06 with reprex v2.0.2}

---

Data:

df <- data.frame(V1 = c(NA, NA, NA), V2 = c(1, 2, NA), V3 = c(NA, NA, NA), V4 = c(2, 2, 1))

df
#>   V1 V2 V3 V4
#> 1 NA  1 NA  2
#> 2 NA  2 NA  2
#> 3 NA NA NA  1