Paste together columns but ignore NAs

user_9974638 0
Viewed 152

I want to paste together multiple columns but ignore NAs.

Here's a basic working example of what the df looks like and what I'd like it to look like. Does anyone have any tips?

df <- data.frame("col1" = c("A", NA, "B", "C"),
                 "col2" = c(NA, NA, NA, "E"),
                 "col3" = c(NA, "D", NA, NA),
                 "col4" = c(NA, NA, NA, NA))

df_fixed <- data.frame("col" = c("A", "D", "B", "C,E"))
paste dataframe r
4 Answers

We may use unite which can have na.rm as argument

library(tidyr)
library(dplyr)
df %>% 
   unite(col, everything(), na.rm = TRUE, sep=",")

-output

  col
1   A
2   D
3   B
4 C,E

Or using base R with do.call and trimws

data.frame(col = trimws(do.call(paste, c(df, sep = ",")),
      whitespace = "(?:,?NA,?)+"))

-output

  col
1   A
2   D
3   B
4 C,E
user_37322710

Using paste.

data.frame(col1=sapply(apply(df, 1, \(x) x[!is.na(x)]), paste, collapse=','))
#   col1
# 1    A
# 2    D
# 3    B
# 4  C,E

Or without apply:

data.frame(col1=unname(as.list(as.data.frame(t(df))) |>
             (\(x) sapply(x, \(x) paste(x[!is.na(x)], collapse=',')))()))
#   col1
# 1    A
# 2    D
# 3    B
# 4  C,E

To add as a column use transform.

transform(df, colX=sapply(apply(df, 1, \(x) x[!is.na(x)]), paste, collapse=','))
#   col1 col2 col3 col4 colX
# 1    A <NA> <NA>   NA    A
# 2 <NA> <NA>    D   NA    D
# 3    B <NA> <NA>   NA    B
# 4    C    E <NA>   NA  C,E

Note: Actually, you also could replace \(x) x[!is.na(x)] by na.omit, since it's attributes vanish; see e.g. @ G. Grothendieck's answer.

user_65740380

A possible base R solution:

df2 <- data.frame(col=apply(df,1, function(x) paste0(na.omit(x), collapse = ",")))

df2

#>   col
#> 1   A
#> 2   D
#> 3   B
#> 4 C,E
user_115644870

Use na.omit and toString. No packages are used.

data.frame(col = apply(df, 1, function(x) toString(na.omit(x)))
##    col
## 1    A
## 2    D
## 3    B
## 4 C, E

Use one of these instead of the anonymous function shown if spaces in the output are a problem:

function(x) paste(na.omit(x), collapse = ",") 
function(x) gsub(", ", ",", toString(na.omit(x)))
user_5165480
Related