r - aggregate multiple rows in dataframe -

i have dataframe looks this:

id=c(3, 3, 4, 5, 5) a_2015 =c("abc", na, na, "abc", na) a_2016 = c("na", "def", "abc", na, "abc") df = data.frame(id, a_2015, a_2016)  df       id    a_2015    a_2016 1    3     abc       na    2    3     na        def 3    4     na        abc        4    5     abc       na 5    5     na        abc 

that means if in column a_2015 entry there na in a_2016 or viceversa. can never have in same row valid entry in both columns a_2015 , a_2016.

i aggregate dataframe like

id    a_2015    a_2016 3     abc       def 4     na        abc 5     abc       abc 

i tried solve aggregate think need apply, or? thankful hints!

you can use dplyr well:

library(tidyverse) df %>%    group_by(id) %>%    summarise(tmp=paste(a_2015, a_2016, collapse = "")) %>%    mutate(tmp=gsub("nana ", "", tmp)) %>%    separate(tmp, = c("a_2015", "a_2016"), sep = " ") # tibble: 3 x 3      id a_2015 a_2016 * <dbl>  <chr>  <chr> 1     3    abc    def 2     4     na    abc 3     5    abc    abc 

or base r:

aggregate(df[,-1],  list(df$id), function(x) gsub("na", "", paste0(x, collapse = "")))   group.1 a_2015 a_2016 1       3    abc    def 2       4           abc 3       5    abc    abc 

then have replace "" na , edit colnames.