我认为以下方法可行。向量化可能是可能的，但我将使用一个循环来处理tryCatch。此外，它的设置允许list作为输入，因为列表中可以有不同的数据类型。该示例将用于平方根，它在非数字时返回错误，在负数时返回警告。它将附加到传递的列表/向量。

squareRoot <- function(x, ot, msg) {
  for (i in seq_along(x)) {
    xsq <- tryCatch(sqrt(x[[i]]),
                    error = function(cond) simpleError(trimws(cond$message)),
                    warning = function(cond) simpleWarning(trimws(cond$message))
    )
    
    if (inherits(xsq, "simpleError")) {
      ot <- c(ot, list(NA_real_))
      msg <- c(msg, "Error Detected")
    } else if (inherits(xsq, "simpleWarning")) {
      ot <- c(ot, list(NA_real_))
      msg <- c(msg, "Warning Detected")
    } else {
      ot <- c(ot, list(xsq))
      msg <- c(msg, "Run Successfully")
    }
  }
  
  list(ot = ot, msg = msg)
}

字符串
现在测试：

x <- list(5, "5", -5, "-5", "Test")
y <- list("A", 36, 81, NA_real_)

Out <- squareRoot(x, list(), c())

Out
$ot
$ot[[1]]
[1] 2.236068

$ot[[2]]
[1] NA

$ot[[3]]
[1] NA

$ot[[4]]
[1] NA

$ot[[5]]
[1] NA

$msg
[1] "Run Successfully" "Error Detected"   "Warning Detected" "Error Detected"  
[5] "Error Detected"

Out <- squareRoot(y, Out$ot, Out$msg)

Out
$ot
$ot[[1]]
[1] 2.236068

$ot[[2]]
[1] NA

$ot[[3]]
[1] NA

$ot[[4]]
[1] NA

$ot[[5]]
[1] NA

$ot[[6]]
[1] NA

$ot[[7]]
[1] 6

$ot[[8]]
[1] 9

$ot[[9]]
[1] NA

$msg
[1] "Run Successfully" "Error Detected"   "Warning Detected" "Error Detected"  
[5] "Error Detected"   "Error Detected"   "Run Successfully" "Run Successfully"
[9] "Run Successfully"

型