CODEWITHSUNDEEP
rmatrixmatrix-multiplication
Kev W.
Kev W.

Reputation: 169

R - Multiply the entire column in R to a list, based on a factor column

Suppose we have the following data-set:

dat<-data.frame(num = 20:29, names = c(rep("Harry",2), rep("Gary",2), rep("Dairy",3), rep("Harry", 3)))

   num names
1   20 Harry
2   21 Harry
3   22  Gary
4   23  Gary
5   24 Dairy
6   25 Dairy
7   26 Dairy
8   27 Harry
9   28 Harry
10  29 Harry

And we also have the following values for each factor level:

fvals <- c(Harry = 1, Gary = 2, Dairy = 3)

The goal is to multiply the num level by these factors by the fvals according to the names level (matching it in the fvals variable). For example, the desired output for this set of data should be

20 # 20 * 1
21 # 21 * 1
44 # 22 * 2
46 # 23 * 2
72 # 24 * 3 
75 # 25 * 3
78 # 26 * 3
27 # 27 * 1
28 # 28 * 1
29 # 29 * 1

I have been doing this by converting the factor variable into a matrix with binary variables for each level, and then proceeding with matrix multiplication. But it was quite confusing trying to convert the matrices / vectors such that R can do multiplication on (and such that the level columns match). Also I'm not sure if the mat mul method will be efficient with a large number of observations. Just wondering if there's a better alternative to do this.

Upvotes: 0

Views: 949

Answers (1)

www
www

Reputation: 39154

Here is an idea. Notice that I set stringsAsFactors = FALSE because it is easier to work with character vector directly.

dat<-data.frame(num = 20:29, 
                names = c(rep("Harry",2), rep("Gary",2), rep("Dairy",3), rep("Harry", 3)),
                stringsAsFactors = FALSE)

fvals <- c(Harry = 1, Gary = 2, Dairy = 3)

dat$num * fvals[dat$names]
# Harry Harry  Gary  Gary Dairy Dairy Dairy Harry Harry Harry 
#    20    21    44    46    72    75    78    27    28    29

Upvotes: 2

Related Questions