CODEWITHSUNDEEP
rstringdata.tablepaste
Chris
Chris

Reputation: 1237

Paste multiple data.table columns into single column based on unique values

I have a data.table that looks like this:

require("data.table")

dt1 <- data.table(VAR1 = c("Brick","Sand","Concrete","Stone"), VAR2 = c(100,23,76,43), VAR3 = c("Place","Location","Place","Vista"), VAR4 = c("Place","Tree","Wood","Vista"), VAR5 = c("Place","Tree","Wood","Forest"))

I would like to paste named columns (my real data has additional columns) together in this order: VAR2, VAR1, VAR3, VAR4 and VAR5. However, I have two conditions:

My expected output would look like this:

dt2 <- data.table(VAR6 = c("100 Brick, Place","23 Sand, Location, Tree","76 Concrete, Place, Wood","43 Stone, Vista, Forest"))

Upvotes: 3

Views: 877

Answers (1)

akrun
akrun

Reputation: 887901

We can use do.call(paste after selecting the column in the order in .SDcols, removve the duplicate words with a regex expression

dt1[,  .(VAR6 = sub(",", " ", gsub("\\b(\\w+)\\b\\s*,\\s*(?=.*\\1)", "", 
      do.call(paste, c(.SD, sep=",")), perl = TRUE))), 
           .SDcols = names(dt1)[c(2:1, 3:5)]]
#                     VAR6
#1:        100 Brick,Place
#2:  23 Sand,Location,Tree
#3: 76 Concrete,Place,Wood
#4:  43 Stone,Vista,Forest

or group by the sequence of rows and do the paste

V6 <- dt1[, sprintf("%s %s, %s", VAR2, VAR1, 
   toString(unique(unlist(.SD)))), 1:nrow(dt1), .SDcols = VAR3:VAR5]$V1
data.table(V6)
#                     V6
#1:         100 Brick, Place
#2:  23 Sand, Location, Tree
#3: 76 Concrete, Place, Wood
#4:  43 Stone, Vista, Forest

Upvotes: 4

Related Questions