通过应用多列函数聚合表

df x1 x2 注意到数值测量 y1 y2 y3

df <- data.frame(x1=sample(letters[1:3], 20, replace=TRUE),
           x2=sample(letters[4:6], 20, replace=TRUE),
           y1=rnorm(20), y2=rnorm(20), y3=rnorm(20))

y x 是一个包含3列的表,该表应输出一个新列:

f <- function(y){   sum((y[,1] - y[,2]) / y[,3]) }

aggregate ,请 dplyr summarizeBy

aggregate(data = df, y1 + y2 + y3 ~ x1 + x2, FUN = f)

为了澄清这一点,预期结果可以通过以下方式获得:

groups <- unique(df[,c("x1", "x2")]) # coocurences of explanatory variables
res <- c()
for (i in 1:nrow(groups)){ # get the subtables
  temp <- df[df$x1 == groups[i,1] & df$x2 == groups[i,2], c("y1", "y2", "y3")]
  res <- c(res, f(temp)) # apply function on subtables
}
groups$res <- res # aggregate results