r - check if every column is na

Question

I have a list of columns within a dataframe which where i want to check if all those columns are NA and create a new column which tells me if they are NA or not.

Here is an example of it working with one column, where Any_Flag is my new column:

ItemStats_2014$Any_Flag <- ifelse(is.na(ItemStats_2014$Item_Flag_A), "Y", "N")

When i try to run the check over multiple columns, I am getting what i expect:

ItemStats_2014$Any_Flag <- ifelse(all(is.na(ItemStats_2014[ ,grep("Flag", names(ItemStats_2014), value = T)])), "Y", "N")

It returns everything to be false or "N".

Answer 1

Data

set.seed(1)
data <- c(LETTERS, NA)
df <- data.frame(Flag_A = sample(data), Flag_B = sample(data), 
                 C = sample(data), D = sample(data), Flag_E = sample(data))

df <- rbind(NA, df)

Code

Identifying all NAs per row:

> df$All_NA <- apply(df[, grep("Flag", names(df))], 1, function(x) all(is.na(x)))
> head(df)
  Flag_A Flag_B    C    D Flag_E All_NA
1   <NA>   <NA> <NA> <NA>   <NA>   TRUE
2      H      K    B    T      Y  FALSE
3      J      W    C    K      P  FALSE
4      O      I    H    I   <NA>  FALSE
5      V      L    M    S      R  FALSE
6      E      N    P    E      I  FALSE

Identifying at least one NA per row:

> df$Any_NA <- apply(df[, grep("Flag", names(df))], 1, function(x) anyNA(x))
> head(df)
  Flag_A Flag_B    C    D Flag_E Any_NA
1   <NA>   <NA> <NA> <NA>   <NA>   TRUE
2      H      K    B    T      Y  FALSE
3      J      W    C    K      P  FALSE
4      O      I    H    I   <NA>   TRUE
5      V      L    M    S      R  FALSE
6      E      N    P    E      I  FALSE

Answer 2

And a data.table way without any apply is:

library(arsenal)
library(data.table)

# dummy data
set.seed(1)
data = c(LETTERS, NA)
dt = data.table(Flag_A=sample(data), Flag_B = sample(data), C=sample(data), D=sample(data), Flag_E=sample(data))
dt = rbind(NA, dt)

# All-NA/Any-NA check
columns_to_check = names(dt)[grep('Flag', names(dt))]
dt[, AllNA:=allNA(.SD), by=1:nrow(dt), .SDcols = columns_to_check]
dt[, AnyNA:=anyNA(.SD), by=1:nrow(dt), .SDcols = columns_to_check]