Aggregating the value of one variable for each class and each ID

Question

I have the following data frame:

id <- c(1,1,1,1,1,2,2,2,2)
spent <- c(10,10,20,10,10,5,5,5,20)
period <- c("f","c","c","v","v","f","c","c","v")
mean.spent <- c(10,15,15,10,10,5,5,5,20)
df <- data.frame(id,spent,period,mean.spent)

what I want is to aggregate the mean spent for each Id in each period as follow:

id  f  c  v    
1   10 15 10
2   5  5  20

Can you please help me to do this?

Answer 1

Use xtabs() along with aggregate() as follows:

df <- data.frame(id = c(1,1,1,1,1,2,2,2,2),
                 spent = c(10,10,20,10,10,5,5,5,20),
                 period = c("f","c","c","v","v","f","c","c","v"),
                 mean.spent = c(10,15,15,10,10,5,5,5,20))

xtabs(spent ~ id + period, aggregate(spent ~ id + period, df, mean))
#    period
# id   c  f  v
#   1 15 10 10
#   2  5  5 20

aggregate calculates the mean per group (as grouped by "id" and "period"), and xtabs does the transformation into this wider format.

Here's how to make it into a data.frame :

temp1 <- xtabs(spent ~ id + period, 
               aggregate(spent ~ id + period, df, mean))
data.frame(id = dimnames(temp1)$id, as.data.frame.matrix(temp1))
#   id  c  f  v
# 1  1 15 10 10
# 2  2  5  5 20

Update: a more direct approach

I always forget about tapply , but this example is a good candidate for when it is convenient.

tapply(df$spent, list(df$id, df$period), mean)
#    c  f  v
# 1 15 10 10
# 2  5  5 20