Data Table - Select Value of Column by Name From Another Column

Question

I have a data table with a number of columns containing values. I have another column which defines which one of those columns whose value I need to select. I am having tr

Answer 1

You can use matrix indexing to pull values from the first and second columns:

mx.idx <- d[, cbind(1:nrow(d), match(name.of.col, names(d)))]
d[, 
  value.of.col:=
    as.matrix(d[, 1:2])[mx.idx]
 ]
d
#    value.1 value.2 name.of.col value.of.col
# 1:     one     two     value.1          one
# 2:     uno     dos     value.2          dos
# 3:       1       2     value.1            1

Answer 2

The following should be memory efficient and a little easier to read/follow.

for (i in unique(d[["name.of.col"]]))
    d[ name.of.col==i, value.of.col:=get(i) ]

d
   value.1 value.2 name.of.col value.of.col
1:     one     two     value.1          one
2:     uno     dos     value.2          dos
3:       1       2     value.1            1

Answer 3

Another option:

d[ , value.of.col := diag(as.matrix(.SD)), .SDcols = d[ , name.of.col]]
> d
   value.1 value.2 name.of.col value.of.col
1:     one     two     value.1          one
2:     uno     dos     value.2          dos
3:       1       2     value.1            1

EDIT add a faster solution:

d[ , value.of.col :=
      melt(d,id.vars='name.of.col')[name.of.col==variable, value]]