combine rows in data frame containing NA to make complete row

Question

I know this is a duplicate Q but I can\'t seem to find the post again

Using the following data

df <- data.frame(A=c(1,1,2,2),B=c(NA,2,NA,4),C=c(3,         


        

Answer 1

This is functionally identical to @Oriol Mirosa's answer without requiring a custom function:

EDIT: NAs must be omitted as per @thelatemail's comment. This answer was also given by @MrFlick in the duplicate thread linked above.

df %>% group_by(A) %>% summarise_all(~first(na.omit(.)))

I wanted to add to this as it seems to come up regularly for me and I've revisited this thread many times. @Oriol Mirosa's answer works, however I'm resistant to it because it's just complex enough to be difficult to remember (hence my return to this thread).

Personally, I also don't like writing small custom functions like if I don't need to. Attempting to substitute coalesce_by_column with the actual coalesce call results in type errors (which I find strange as the rows aren't interacting with each other but whatever). This can be resolved by first doing mutate_all(as.character), however my goal here is to minimize syntax so it's easily remembered on the fly.

Furthermore, this substitution changes the behavior such that non-identical values within a column throws an error (why things sometimes behave slightly differently within a function is beyond me). This behavior may be preferred in some situations, however in that case I would recommend @Jerry T's solution as there is no custom function and the ones used are familiar, readable, and the ordering of them (na.omit and unique) isn't relevant.