CODEWITHSUNDEEP
rmergeconcatenation
Harry Palmer
Harry Palmer

Reputation: 478

R: Merge of rows in same data table, concatenating certain columns

I have my data table in R. I want to merge rows which have an identical customerID, and then concatenate the elements of other merged columns.

I want to go from this:

   title  author customerID
1 title1 author1          1
2 title2 author2          2
3 title3 author3          1

to this:

           title           author Group.1
1 title1, title3 author1, author3       1
2         title2          author2       2

Upvotes: 11

Views: 14676

Answers (2)

A5C1D2H2I1M1N2O1R2T1
A5C1D2H2I1M1N2O1R2T1

Reputation: 193517

The aggregate function should help you in finding a solution:

dat = data.frame(title = c("title1", "title2", "title3"),
                 author = c("author1", "author2", "author3"),
                 customerID = c(1, 2, 1))
aggregate(dat[-3], by=list(dat$customerID), c)
#   Group.1 title author
# 1       1  1, 3   1, 3
# 2       2     2      2

Or, just make sure you add stringsAsFactors = FALSE when you are creating your data frame and you're pretty much good to go. If your data are already factored, you can use something like dat[c(1, 2)] = apply(dat[-3], 2, as.character) to convert them to character first, then:

aggregate(dat[-3], by=list(dat$customerID), c)
#   Group.1          title           author
# 1       1 title1, title3 author1, author3
# 2       2         title2          author2

Upvotes: 14

sgibb
sgibb

Reputation: 25736

Maybe not the best solution but easy to understand:

df <- data.frame(author=LETTERS[1:5], title=LETTERS[1:5], id=c(1, 2, 1, 2, 3), stringsAsFactors=FALSE)

uniqueIds <- unique(df$id)

mergedDf <- df[1:length(uniqueIds),]

for (i in seq(along=uniqueIds)) {
    mergedDf[i, "id"] <- uniqueIds[i]
    mergedDf[i, "author"] <- paste(df[df$id == uniqueIds[i], "author"], collapse=",")
    mergedDf[i, "title"] <- paste(df[df$id == uniqueIds[i], "title"], collapse=",")
}

mergedDf
#  author title id
#1    A,C   A,C  1
#2    B,D   B,D  2
#3      E     E  3

Upvotes: 3

Related Questions