如何在 R 函数之间传递参数？

Question

我有 2 个小函数（DoAnalysis() 和 ProcessRes()）和 1 个最终函数（PerformSim()），它们将调用 2 个小函数。下面列出了详细的示例函数。

每个函数都包含一个唯一的参数（非重叠参数）：

• 函数PerformSim() 包含1 个唯一参数nsim；
• 函数DoAnalysis() 包含1 个唯一参数obj；
• 函数 ProcessRes() 包含 1 个唯一参数 results

对于剩余的参数，函数DoAnalysis()有完整的集合，函数DoAnalysis()包含函数DoAnalysis()的参数子集，因此，函数PerformSim()需要包含调用这两个函数@的信息987654334@和ProcessRes()。

以下代码不起作用，因为我不确定如何在函数之间正确传递参数。

请帮助并提前非常感谢您的帮助。

library(purrr)
DoAnalysis <- function(obj, nS1 = 40, nS2 = 40, nS3 = 20, cutoff) {

    cumS2 <- nS1 + nS2      ## this is the total number of patients at the end of Stage 2
    nmax  <- cumS2 + nS3    ## this is the total number of patients at the end of study

    ## compute number of events for 1st stage
    Res_n <- sum(obj[1:nS1]) ## this is the number of response in Stage 1

    ## compute number of events for 2nd stage
    Res_n2 <- sum(obj[1:cumS2]) ## this is the number of response in Stage 2

    ### Now, it is the final analysis
    Res_F <- sum(obj[1:nmax]) ## this is the number of response in Final

    list(Res_n, Res_n2, Res_F)
}

ProcessRes <- function(results, cutoff){
  part1 <- part2 <- part3 <- c()

  for (i in 1:length(results)){
     part1 <- as.vector(c(part1, results[[i]][[1]]))
     part2 <- as.vector(c(part2, results[[i]][[2]]))
  }
  resultsTypeI <- as.data.frame(cbind(part1, part2, part3))

  names(resultsTypeI) <- c("Res_n", "Res_n2")

  resultsTypeI$fulFlag <- ifelse(resultsTypeI$Res_n2 < cutoff, 1, 0)
  R_1 <- mean(resultsTypeI$fulFlag)   

  return(c(R_1))
}

set.seed(20201022)

PerformSim <- function(nsim, nS1 = 40, nS2 = 40, nS3 = 20, cutoff = 26){
  ## this is the simulation  
  total <- nS1 + nS2 + nS3
  SimuTypeI <- map(1:nsim, ~rbinom(total, 1, 0.4))
  results <- map(SimuTypeI, ~DoAnalysis(.x))
  ProcessRes(results = results)
}
PerformSim(nsim = 1000)

Answer 1

您的所有函数定义和参数看起来都不错，但在您的 PerformSim() 函数定义中，您必须在 ProcessRes() 中传递参数 cutoff。像这样。

PerformSim <- function(nsim, nS1 = 40, nS2 = 40, nS3 = 20, cutoff = 26){
  ## this is the simulation  
  total <- nS1 + nS2 + nS3
  SimuTypeI <- map(1:nsim, ~rbinom(total, 1, 0.4))
  results <- map(SimuTypeI, ~DoAnalysis(.x))
  ProcessRes(results = results, cutoff)
}