How to implement coalesce efficiently in R

Question

Background

Several SQL languages (I mostly use postgreSQL) have a function called coalesce which returns the first non null column element for each row. This can b

Answer 1

Here is my solution:

coalesce <- function(x){ y <- head( x[is.na(x) == F] , 1) return(y) } It returns first vaule which is not NA and it works on data.table, for example if you want to use coalesce on few columns and these column names are in vector of strings:

column_names <- c("col1", "col2", "col3")

how to use:

ranking[, coalesce_column := coalesce( mget(column_names) ), by = 1:nrow(ranking)]