I have a data frame of three columns:title, text and grp.

The data frame looks like this:

I want to add the comments as columns to this dataframe and keep the value as its text corresponding to each title.

desired dataframe:

I tried re-casting:

library(reshape2)

recast(df, text ~ comment, id.var = c("text"))

But gives me an error:

Error in unique.default(x) : unique() applies only to vectors

CodePudding user response：

Try this

library(dplyr , warn.conflicts = FALSE)

cdf <- df %>% filter(title == "comments") %>%  select( text, grp) 

colnames(cdf)[1] <- "comments"

nocdf <- df %>% select(title , text , grp) %>% filter(title != "comments")

new_df <- right_join(nocdf , cdf , by = "grp")

new_df %>% relocate(grp , .after = last_col())
#>        title                            text
#> 1     Week.2       In.Week.2.the.encoding...
#> 2 Statistics Statistics.is.the.discipline...
#>                                  comments grp
#> 1     collection.of.comments.about.Week.2   2
#> 2 collection.of.comments.about.Statistics   3

^{Created on 2022-06-05 by the reprex package (v2.0.1)}

CodePudding user response：

You could try to split every two rows and cbind.

split(dat, 1:2) |> {\(.) cbind(.$`1`, comments=.$`2`[, 2])}()
#        title                            text grp                                comments
# 1     Week 2       In Week 2 the encoding...   2     collection of comments about Week 2
# 3 Statistics Statistics is the discipline...   3 collection of comments about Statistics

Data:

dat <- structure(list(title = c("Week 2", "comments", "Statistics", 
"comments"), text = c("In Week 2 the encoding...", "collection of comments about Week 2", 
"Statistics is the discipline...", "collection of comments about Statistics"
), grp = c(2L, 2L, 3L, 3L)), class = "data.frame", row.names = c(NA, 
-4L))