data.table“ sumproduct”样式矢量乘法 [英] data.table "sumproduct" style vector multiplication

问题描述

在这个玩具示例中，我想用每行各自的值求和一个系数列表，并将结果分配给新列。下面的代码适用于给定的记录，但是当我删除i参数时，它对我而言表现异常。我可以循环执行或应用，但似乎缺少一种data.table方式。

In this toy example, I want to "sumproduct" a list of coefficients with each row's respective value and assign the result to a new column. The code below works for a given record, but when I remove the i parameter it behaves unexpectedly to me. I could do this in a loop or apply, but it seems like there's a data.table way that I'm missing.

DT <- data.table(mtcars)
vars <- c("mpg","cyl","wt")
coeffs <- c(2,3,4)
DT[1,Calc := sum(coeffs*DT[1,vars,with=FALSE])]  # row 1 is assigned 70.480
DT[,Calc := sum(coeffs*DT[,vars,with=FALSE])]  # all rows assigned 2830.416

推荐答案

使用矩阵乘法：

coeffs <- as.vector(c(2,3,4))
dt2 <- DT[,Calc := as.matrix(DT[,..vars])%*%coeffs]

这篇关于data.table“ sumproduct”样式矢量乘法的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！