How can one transform a data table using data.table

Question

I would like to transform following data table

da<-data.table(Gr=c("A","B","B","B","A","B"),
   mmoSze=c("SCRSCR","SCRSCR","OSNOSN","Eig.SCRSCR","Eig.SCRSCR","Eig.OSNOSN"),
              SCR=c(25, 43, 61, 79, 97, 115))

into:

da.goal<-data.table(GR=c(   "A",    "B"),
SCRSCR=c(   25, 43),
OSNOSN=c(   0,  61),
Eig.SCRSCR=c(   97, 79),
Eig.OSNOSN=c(   0,  115))

using data.table! One possible undesirable way would be to divide the table and then merge them. What is the elegant way to do it?

you have the `melt` function to transform from wide to long format, and `dcast` to transform from long to wide. See https://cran.r-project.org/web/packages/data.table/vignettes/datatable-reshape.html — denis, Jul 05 '19 at 21:11

score 2 · Accepted Answer · answered Jul 05 '19 at 20:59

2

dcast() is a data.table function, and I think this is elegant enough:

dcast(da, Gr ~ mmoSze)

Only problem is that you get NA instead of 0, but then you can do:

res = dcast(da, Gr ~ mmoSze)
res[is.na(res)] <- 0

answered Jul 05 '19 at 20:59

Fino

1,774
11
21

score 1 · Answer 2 · answered Jul 05 '19 at 21:22

1

You need to convert your long table to wide:

dcast(da, Gr ~ mmoSze, fill = 0)

fill = 0 will put a zero where it can't find a value.

answered Jul 05 '19 at 21:22

PavoDive

6,322
2
29
55

How can one transform a data table using data.table

2 Answers2