Reshaping data frame with duplicates

Question

I have what should be a simple reshaping problem, but I can't figure it out. Part of my data looks like this:

foo <- structure(list(grade = c(3, 3, 4, 4, 5, 5, 6, 6, 7, 7, 8, 8, 
3, 3, 4, 4, 5, 5, 6, 6), var.type = structure(c(3L, 2L, 3L, 2L, 
3L, 2L, 3L, 2L, 3L, 2L, 3L, 2L, 3L, 2L, 3L, 2L, 3L, 2L, 3L, 2L
), .Label = c("Raw Score", "SE", "SS"), class = "factor"), var.val = c(120L, 
47L, 120L, 46L, 120L, 46L, 120L, 47L, 120L, 46L, 120L, 46L, 120L, 
12L, 120L, 14L, 120L, 16L, 120L, 20L)), .Names = c("grade", "var.type", 
"var.val"), row.names = c(2L, 3L, 4L, 5L, 6L, 7L, 8L, 9L, 10L, 
11L, 12L, 13L, 15L, 16L, 17L, 18L, 19L, 20L, 21L, 22L), class = "data.frame")

or

  grade var.type var.val
2     3       SS     120
3     3       SE      47
4     4       SS     120
5     4       SE      46
6     5       SS     120
7     5       SE      46

I want to make it look like this:

grade    SS    SE
3        120   47
4        120   46
5        120   46

and so on. I have tried reshape, cast, and dcast as in this thread:

Reshaping dataset

but nothing seems to work. I would really appreciate some help. TIA.

Answer 1

library(plyr)
library(reshape2)
# First we add a grouping variable to deal with the duplicates
foo <- ddply(foo, .(grade, var.type), function(x) { x$group <- 1:nrow(x); x })
dcast(foo, grade + group ~ var.type, value.var= "var.val")[-2]

 grade SE  SS
1      3 47 120
2      3 12 120
3      4 46 120
4      4 14 120
5      5 46 120
6      5 16 120
7      6 47 120
8      6 20 120
9      7 46 120
10     8 46 120

Answer 2

If you don't have any duplicates, this will work nicely:

ss <- subset(foo, var.type=='SS')
se <- subset(foo, var.type=='SE')
ss <- data.frame(grade=ss$grade,SS=ss$var.val)
se <- data.frame(grade=se$grade,SE=se$var.val)
bar <- merge(ss,se,by='grade')

Answer 3

If you want to reshape and you have duplicates, you're going to need to give each pair a unique id:

foorle <- rle(foo$grade)
fooids <- rep(seq_len(length(foorle$values)), times=foorle$lengths)

fooids
 [1]  1  1  2  2  3  3  4  4  5  5  6  6  7  7  8  8  9  9 10 10

Now you'll be able to use reshape properly:

idfoo <- cbind(id=fooids, foo)

library(reshape)
dcast(idfoo, id+grade~var.type, value.var="var.val")

   id grade SE  SS
1   1     3 47 120
2   2     4 46 120
3   3     5 46 120
4   4     6 47 120
5   5     7 46 120
6   6     8 46 120
7   7     3 12 120
8   8     4 14 120
9   9     5 16 120
10 10     6 20 120

EDIT: Please note I'm assuming your data is in order, else you'll have problems distinguishing between duplicates. If it isn't, you can always use order so that it is.

Answer 4

It is not as pretty as reshape, but

data.frame(grade = foo[2 * (1:(nrow(foo)/2)),]$grade, 
           SS =  foo[foo$var.type == "SS", ]$var.val, 
           SE =  foo[foo$var.type == "SE", ]$var.val ) 

produces

   grade  SS SE
1      3 120 47
2      4 120 46
3      5 120 46
4      6 120 47
5      7 120 46
6      8 120 46
7      3 120 12
8      4 120 14
9      5 120 16
10     6 120 20

You have to assume the data comes in pairs of rows for this.