Is there a R function to remove columns that only have missing values?

Question

I have around more than 1000 columns in my data frame that only has NA values throughout. Is there a function in R which could remove columns that have NA values throughout?

Can't find a duplicate. Here's a **self promotion** comment: `mde::drop_na_if(airquality,sign="eq",percent_na=100)` There exist simpler ways like `complete.cases`,`drop_na`(`tidyr`), etc from [mde](https://www.github.com/Nelson-Gon/mde) that I wrote. — NelsonGon, Mar 26 '20 at 11:52

user2974951 · Accepted Answer · 2020-03-26T12:07:09.247

0

Not one function but

df[,colSums(apply(df,2,is.na))<nrow(df)]

edited Mar 26 '20 at 12:07

answered Mar 26 '20 at 11:41

user2974951

9,535
1
17
24

2

You can avoid `apply()` by using `df[colMeans(is.na(df)) != 1]`. – tmfmnk Mar 26 '20 at 11:54

r.user.05apr · Answer 2 · 2020-03-26T11:51:23.470

0

d <- data.frame(a = c(NA, rep(1, 9)),
                b = rep(NA, 10))

d[vapply(d, function(x) all(!is.na(x)), logical(1))]

edited Mar 26 '20 at 11:51

answered Mar 26 '20 at 11:43

r.user.05apr

5,356
3
22
39

Why is `a` dropped? – NelsonGon Mar 26 '20 at 11:57
2

Although this code might solve the problem, a good answer should also explain what it does and how it helps? – Suraj Kumar Mar 26 '20 at 12:44

Is there a R function to remove columns that only have missing values?

2 Answers2