Example Data
Consider the two following data.frames as an example
set.seed(12)
d1 = cbind(
expand.grid(1:4,LETTERS[1:3]),
rnorm(12,10,1)
)[sample(1:12),]
d2 = as.data.frame(
cbind(
sample(c(1,2,4),900, replace=TRUE),
sample(LETTERS[1:3],900,replace=TRUE)
)[sample(1:900),]
)
names(d1) = c("x","y","z")
names(d2) = c("x","y")
d1 is small and contain all possible combinations of x and y and contains the z variable. d2 is much longer, does not necessarily contain all possible combinations of x and y and does not contain the z variable. In both variables, there is no logical order for the combinations of x and y (as in the example).
Goal
How can I enter the values of the z variable from d1 into d2 respecting the combination of x and y?
Bad solution
d2$z = 0
for (x in unique(d1$x))
{
for (y in unique(d1$y))
{
d2$z[d2$x == x & d2$y == y] = d1$z[d1$x == x & d1$y == y]
}
}
head(d2)
x y z
1 2 B 9.727704
2 1 C 9.893536
3 2 A 11.577169
4 1 A 8.519432
5 4 C 8.706118
6 2 B 9.727704
However, it is terribly slow when the data.frames are a few millions rows long. Also, it is not very flexible.
Is there a better alternative?
