Multiplication in FUN argument-CodePudding

I have this dataframe

x <- data.frame(
        matrix(
          c(letters[1:3], c("x", "x", "y") ,
          sample(c(rep(1,100),0), size = 1),
          sample(c(rep(1,100),0), size = 1),
          sample(c(rep(1,100),0), size = 1)), ncol = 3)
)

I would like to do multiplication by group X and Y.

My suggestion

agg <- aggregate(x$X3,
                 by = list(x$X2),
                 FUN = *)

I would like to use something like sum, mean byt to multiply

CodePudding user response：

is to sum as * is to prod (for product).

Your sample data follows the anti-pattern of data.frame(matrix()). A matrix can only have one data type. You mix character and numeric data in the matrix, and the matrix makes it all character class, and you can't do math on characters. Here's proper sample data and a demonstration the solution works. Also note that using by = X["X2"] instead of by = list(x$X2) gives a nicer column name in the result.

(x <- data.frame(
          X1 = letters[1:3],
          X2 = c("x", "x", "y") ,
          X3 = 2:4
))
#   X1 X2 X3
# 1  a  x  2
# 2  b  x  3
# 3  c  y  4

aggregate(x$X3, by = x["X2"], FUN = prod)
#   X2 x
# 1  x 6
# 2  y 4

CodePudding user response：

Either use prod or use Reduce with *. Also convert X3 to numeric and and use single brackets as shown to preserve the names. Alternately use the aggregate formula method, shown only for prod but applies to Reduce as well.

xx <- transform(x, X3 = as.numeric(X3))

aggregate(xx["X3"], by = xx["X2"], FUN = prod)

aggregate(xx["X3"], by = xx["X2"], FUN = Reduce, f = `*`) # same

aggregate(X3 ~ X2, xx, FUN = prod)

A better example might be to use mtcars that comes with R:

aggregate(mtcars["mpg"], by = mtcars["cyl"], FUN = prod)

aggregate(mtcars["mpg"], by = mtcars["cyl"], FUN = Reduce, f = `*`) # same

aggregate(mpg ~ cyl, mtcars, FUN = prod)