Question

假设我有一张这样的桌子：

data <- c(1,2,3,6,5,6,9,"LC","LC","HC","HC","LC","HC","ALL")
attr(data,"dim") <- c(7,2)
data
     [,1] [,2] 
[1,] "1"  "LC" 
[2,] "2"  "LC" 
[3,] "3"  "HC" 
[4,] "6"  "HC" 
[5,] "5"  "LC" 
[6,] "6"  "HC" 
[7,] "9"  "ALL"

现在我想操纵数据，看起来像这样：

     [,"LC"] [,"HC"] [,"ALL"] 
[1,] "1"     "3"     "9"
[2,] "2"     "6"
[3,] "5"     "6"

有没有办法在 R 中执行此操作，或者这是不可能的，我应该尝试其他方式来访问我的数据吗？

Answer 1

使用split可以非常接近。这将返回包含所需值的列表，然后您可以使用lapply或任何其他列表操作函数：

split(data[, 1], data[, 2])

$ALL
[1] "9"

$HC
[1] "3" "6" "6"

$LC
[1] "1" "2" "5"

如果你必须以矩阵格式输出，那么我建议你用NA填充短矢量：

x <- split(data[, 1], data[, 2])
n <- max(sapply(x, length))

pad_with_na <- function(x, n, padding=NA){
  c(x, rep(padding, n-length(x)))
}

sapply(x, pad_with_na, n)

这导致：

     ALL HC  LC 
[1,] "9" "3" "1"
[2,] NA  "6" "2"
[3,] NA  "6" "5"

Answer 2

示例数据

我更喜欢将数据读入data.frame，因为它检查向量是否长度相等。

data <- data.frame(X=c(1,2,3,6,5,6,9), 
                   Y=c("LC","LC","HC","HC","LC","HC","ALL"))

CODE

data <- unstack(data, form=X~Y)# easier to read than split
Nmax <- do.call(max, lapply(data,length))
sapply(data, "[", seq(Nmax))# "borrowed" from other answer in SO

创建具有不同行维度的表

2 个答案: