Unstacking a data frame

Question

I have a data frame which looks like this:

d <- c("a", "b", "c", "a", "b", "c", "a", "b", "c")
par <- c("a1", "a2", "a3", "a4", "a5", "a6", "a7", "a8", "a9")
df1 <- as.data.frame(cbind(id, par))

I would like it to look like this:

a <- c("a1", "a4", "a7")
b <- c("a2", "a5", "a8")
c <- c("a3", "a6", "a7")
df2 <- data.frame(rbind(a, b, c))

I have limited experience / understanding of re-arranging data frames, so far I have tried using the functions unstack, dcast and melt. — user4846627, Apr 29 '15 at 13:57

score 1 · Accepted Answer · answered May 23 '15 at 12:47

1

You could do this with unstack

as.data.frame(t(unstack(df1,par~id)))
#  V1 V2 V3
#a a1 a4 a7
#b a2 a5 a8
#c a3 a6 a9

Or using dcast after creating a sequence column for the 'id' group.

library(data.table)#v1.9.5
dcast(setDT(df1)[, ind:= 1:.N, id], id~ind, value.var='par')

answered May 23 '15 at 12:47

akrun

874,273
37
540
662

Thanks, I was trying to do this but not having much luck – user4846627 May 28 '15 at 09:48
@user4846627 Did this help or not? Sorry, I didn't understand what you really meant. Based on the example you provided, I am getting the expected result – akrun May 28 '15 at 09:49
Hi, yes it helped. Thanks for your reply. – user4846627 May 28 '15 at 09:55

score 0 · Answer 2 · answered Apr 29 '15 at 15:41

I have solved this as below, seems too messy though

id <- c("a", "b", "c", "a", "b", "c", "a", "b", "c")
par <- c("a1", "a2", "a3", "a4", "a5", "a6", "a7", "a8", "a9")
df1 <- as.data.frame(cbind(id, par))


ids <- levels(df1$id)

newdf <- list()

for ( i in 1:length(ids)){
tempdf <- df1[df1$id == ids[i],]
id <- tempdf[1,1]
parcels <- tempdf[,2]
row <- cbind(as.character(id), t(as.character(parcels)))
tempdf <- as.data.frame(row)
newdf[[i]] <- tempdf
}

df3 <- do.call(rbind.fill, newdf)

I was using a much larger data frame which resulted in a varying number of columns, hence rbind.fill — user4846627, Apr 29 '15 at 15:46

Unstacking a data frame

2 Answers2