argument 'pattern' has length > 1 and only the first element will be used - GSUB()

Question

I have the following problem.

table <- data.frame(col1 = c("cars1 gm", "cars2 gl"), col2 = c("cars1 motor mel", "cars2 prom del"))

      col1            col2
1 cars1 gm cars1 motor mel
2 cars2 gl  cars2 prom del

table$word <- gsub(table$col1, ' ', table$col2) 

Warning message:  In gsub(table$col1, " ", table$col2) :  argument
'pattern' has length > 1 and only the first element will be used

How do I create a new column called word containing only those values of col2 which do not appear in col1?

      col1            col2       word
1 cars1 gm cars1 motor mel  motor mel
2 cars2 gl  cars2 prom del   prom del

Answer 1

You can use gsub to build your lookup and then sapply over columns to perform the gsub of interest:

table$col1 <- gsub(" ", "|", table$col1)
table$word <- sapply(1:nrow(table), function(x) gsub(table$col1[x], "", table$col2[x]))

table
#      col1            col2       word
#1 cars1|gm cars1 motor mel  motor mel
#2 cars2|gl  cars2 prom del   prom del

---

Using a similar idea as the above answer but using mapply instead of sapply:

table$word <- mapply(function(x, y) gsub( gsub(" ", "|", x), "", y),
                                    table$col1,
                                    table$col2)

Answer 2

You can use mapply,

#Make sure you read your data with stringsAsFactors = FALSE, 
table<-data.frame(col1=c("cars1 gm","cars2 gl"),
                  col2=c("cars1 motor mel", "cars2 prom del"), stringsAsFactors = FALSE)

table$word <- mapply(function(x, y) 
                     trimws(gsub(sapply(strsplit(x, ' '), paste, collapse = '|'), '', y)), 
                     table$col1, table$col2)
table
#      col1            col2      word
#1 cars1 gm cars1 motor mel motor mel
#2 cars2 gl  cars2 prom del  prom del