Question

Possible Duplicate:
split and combine by factor into new columns

Using example data here for easier illustrative purposes.
I have the full greek alphabet as my sample set.
I also have the full latin alphabet.

I’m currently working with a table that looks like this. This would be a small subset. My actual data is in the 100s of thousands of rows.

sample  call    letter
alpha   y       A
alpha   y       D
alpha   n       C
beta    y       F
beta    y       B
gamma   n       F
gamma   y       B
gamma   n       A
delta   m       E
epsilon n       F
epsilon n       C
epsilon n       B
zeta    m       A
zeta    y       D

I would like to make a final table where the greek alphabet are rows and the columns are the latin alphabet and I’d only provide the “call” whenever available, otherwise NA or 0 could be the entry.

End result would be something like so.

        A   B   C   D   E   F
alpha   y   NA  n   y   NA  NA
beta    NA  y   NA  NA  NA  y
gamma   n   y   NA  NA  NA  n
delta   NA  NA  NA  NA  m   NA
epsilon NA  n   n   NA  NA  n
zeta    m   NA  NA  y   NA  NA

Any idea how I could achieve something like this? Even a general pointer into what would be great.

Thanks in advance.

Answer 1

dcast from the reshape2 package does this nicely

library(reshape2)
dat <- structure(list(sample = c("alpha", "alpha", "alpha", "beta", "beta", "gamma", "gamma", "gamma", "delta", "epsilon", "epsilon", "epsilon", "zeta", "zeta"), call = c("y", "y", "n", "y", "y", "n", "y", "n", "m", "n", "n", "n", "m", "y"), letter = c("A", "D", "C", "F", "B", "F", "B", "A", "E", "F", "C", "B", "A", "D")), .Names = c("sample", "call", "letter"), class = "data.frame", row.names = c(NA, -14L))

dcast(dat, sample ~ letter, value.var='call')


   sample    A    B    C    D    E    F
1   alpha    y <NA>    n    y <NA> <NA>
2    beta <NA>    y <NA> <NA> <NA>    y
3   delta <NA> <NA> <NA> <NA>    m <NA>
4 epsilon <NA>    n    n <NA> <NA>    n
5   gamma    n    y <NA> <NA> <NA>    n
6    zeta    m <NA> <NA>    y <NA> <NA>