Count number of rows with at least one value

Question

I have a df like this:

a   b   c   d
1   1   1   1
2   2   3   NA
3   NA  NA  2 
4   NA  2   NA
5   NA  NA  NA

Is it possible to get a count of how many rows doesn't have non missing values on all variables, so for example here it will return 4 as the rest of the rest have at least one none NA value on some of it's variables.

Shouldn't "a count of how many rows have non missing values on all variables" return 1 for your example data (the first row) ? Maybe you need to explain the logic more clearly. — neilfws, Sep 04 '19 at 04:46
Are you neglecting column `a` from calculation? So `sum(rowSums(!is.na(df[-1])) > 0)` ? — Ronak Shah, Sep 04 '19 at 05:03
*"how many rows doesn't have non missing values on all variables"* Do you mean how many rows have at last one `NA` entry? E.g. `sum(is.na(rowSums(df)))`? — Maurits Evers, Sep 04 '19 at 05:22
No. On this example the answer should 4. You need to exclude the a var from the calculations since that one is the ID number — EGM8686, Sep 04 '19 at 12:47
@EGM8686 `sum(is.na(rowSums(df)))` returns 4. So is that what you're after? Your problem statement is not very clear. — Maurits Evers, Sep 04 '19 at 22:10

score 0 · Answer 1 · answered Sep 05 '19 at 05:24

Since it is clear now that we want to ignore column a from calculation. Here are few ways:

Using rowSums

sum(rowSums(!is.na(df[-1])) > 0)
#[1] 4

#OR

sum(rowSums(is.na(df[-1])) != (ncol(df) - 1))

Using apply

sum(apply(!is.na(df[-1]), 1, any))

#OR

sum(!apply(is.na(df[-1]), 1, all))

Using filter_at from dplyr filters the row based on condition but we can use nrow to get the number of rows which satisfies our requirement.

library(dplyr)
df %>%  filter_at(-1, any_vars(!is.na(.))) %>% nrow

Count number of rows with at least one value

1 Answers1