Question

如果我有一个任意长度的字符串向量说;

vec <- c("a","b","c","d","e","f","g","h","i")

我正在寻找一个函数，使用下面显示的策略从该向量中采样n个值。显示数字，因为它很难解释。

function call: 
result:
schematic diagram:

fn(vector=vec,n=1)
"a"

|
a b c d e f g h i


fn(vector=vec,n=2)
"a" "i"
 _______________
|               |
a b c d e f g h i


fn(vector=vec,n=3)
"a" "e" "i"
 _______________
|       |       |
a b c d e f g h i


fn(vector=vec,n=4)
"a" "c" "g" "i"
 _______________
|   |     |     |
a b c d e f g h i


fn(vector=vec,n=5)
"a" "c" "e" "g" "i"
 _______________
|   |   |   |   |
a b c d e f g h i

抽样不一定准确。值可以来自大致正确的区域，但必须一致。字符串向量可以是偶数或奇数。

Answer 1

一种方法是使用seq()，利用它的length.out=参数来获得您寻求的均匀间隔索引：

fn <- function(x, n) {
    x[round(seq(1,length(x), length.out=n))]
}

## Check that it works
fn(vec, 1)
# [1] "a"
fn(vec, 2)
# [1] "a" "i"
fn(vec, 4)
# [1] "a" "d" "f" "i"
fn(vec, 8)
# [1] "a" "b" "c" "d" "f" "g" "h" "i"

Answer 2

这应该可以满足您的需求：

fn <- function(myVector, n) {
  # check for valid n
  if(n > length(myVector)) stop("invalid n")

  if(n == 1) return(myVector[1])
  if(n == 2) return(myVector[c(1, length(myVector))])

  middleSpots <- ceiling(length(vec) * (1:(n-2) / (n-1)))
  return(myVector[c(1, middleSpots, length(myVector))])
}

通过相等分割对矢量进行采样

2 个答案: