使用lapply进行错误处理-输出失败元素的索引 [英] Error handling with lapply -- output the index of failed elements

问题描述

对使用lapply处理错误的疑问总是返回 NA 或 NULL ，当元素失败时，即

  myfun<-函数{tryCatch(doSomething(s)，错误=函数(e){return(NULL)}} 

但是，这还不够通用，因为 doSomething(s)可能会返回 NULL 或 NA 本身.因此，理想情况下，我希望编写 myfun ，以便在 lapply(mylist，myfun)之后，我能够以某种方式获取所有失败元素的索引.该怎么做?

解决方案

使用 identity()

处理并释放错误

  res = lapply(list(1，"two"，3)，function(i)tryCatch({sqrt(i)}，错误=身份) 

检查错误

  vapply(res，是，逻辑(1)，错误") 

返回错误条件通常比返回'magic'值(例如 NA 或 NULL )更好，除非下游分析与返回的值无缝地配合./p>

作为更高级的解决方案，请创建更复杂的条件或扩展错误类别

  my_condition = function(err)结构(列表(消息= conditionMessage(err)，original = err)，class = c("my"，"condition")) 

并退回

  res<-lapply(list(1，"two"，3)，function(i){试着抓({sqrt(i)}，错误= my_condition)}) 

Answer to question about error handling with lapply always return NA or NULL when an element fails, i.e.

myfun <- function(s) {
  tryCatch(doSomething(s), error = function(e) { return(NULL) }
}

However, this is not general enough since doSomething(s) may return NULL or NA itself. Therefore, ideally I want myfun written so that after lapply(mylist, myfun) I can somehow get all the indices of failed elements. How to do this?

解决方案

Catch and release the error by handing it with identity()

res = lapply(list(1, "two", 3), function(i) tryCatch({
    sqrt(i)
}, error=identity)

Check for errors

vapply(res, is, logical(1), "error")

Returning the error condition is often better the returning a 'magic' value like NA or NULL, unless the down-stream analysis works seamlessly with the value returned.

As a more advanced solution, create a more elaborate condition or extend the error class

my_condition = function(err)
    structure(list(message=conditionMessage(err),
                   original=err), class=c("my", "condition"))

and return that

res <- lapply(list(1, "two", 3), function(i) {
    tryCatch({
        sqrt(i)
    }, error=my_condition)
})

这篇关于使用lapply进行错误处理-输出失败元素的索引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！