为何单独运行函数代码行正常，整函数运行却报错？

我编写了一个脚本，通过多个函数将不同国家的数据集标准化为统一格式。批量运行代码时出现错误，但单独执行FormatTMtCensusData函数内的代码后，再运行主函数就能正常工作。

主脚本代码 #

machine <- ajs
database.code <- "NOR"

#1. 导入数据集——此部分运行无问题
orig.data.list <- LoadTMtOriData (database.code, machine = machine, team.qc = F)

#2. 转换数据格式
tmt.data.list <- Conver2TMtData (orig.data.list, database.code = database.code, machine = machine)

报错信息 #

Error in FormatTMtCensusData(tree = tree_test, database.code = "NOR") : 
  object 'census_info' not found

Conver2TMtData函数代码 #

Conver2TMtData <- function (orig.data.list, database.code, machine) {

td1 <- orig.data.list [[1]]; spp1 <- orig.data.list [[2]]; 

if (database.code == 'NOR') {
    tree <- FormatTMtTreeData (td = td1, spp = spp1, md = md1, database.code = 'NOR')
    write.csv(tree, "/Users/ajdinstarcevic/OneDrive - University of Eastern Finland/UEF/WUR/WORK/Mortality model/Data/NOR/tree.csv")
    spp <- FormatTMtSppData (tree, database.code = 'NOR')
    tree_test <- tree
    census_info <- FormatTMtCensusData (tree=tree_test, database.code = 'NOR') 
    
    
  }
}

FormatTMtCensusData函数代码 #

FormatTMtCensusData <- function (tree, database.code) {

if (database.code == 'NOR') {
   census_info <- unique (tree[,c('database.code','tmt.census.id','census.id', 'census.date','census.n','tmt.plot.id','plot.id')])
   census_info$census.contact <- 'NA'
    census_info$census.contact.email <- 'NA'
    
    
 }
}

问题原因 #

FormatTMtCensusData函数内部创建了census_info对象，但没有返回该对象。批量运行时，函数执行完毕后内部环境销毁，census_info不会传递给调用它的Conver2TMtData函数；而单独运行函数内的代码时，census_info会被创建在全局环境中，后续调用Conver2TMtData时能读取到这个全局变量，因此不会报错。

解决方案 #

修改FormatTMtCensusData函数，确保它返回census_info对象。有两种方式：

1. 在函数末尾添加return(census_info)：

FormatTMtCensusData <- function (tree, database.code) {

if (database.code == 'NOR') {
   census_info <- unique (tree[,c('database.code','tmt.census.id','census.id', 'census.date','census.n','tmt.plot.id','plot.id')])
   census_info$census.contact <- 'NA'
    census_info$census.contact.email <- 'NA'
    
 }
 return(census_info)
}

2. 直接将census_info作为函数的最后一行（R函数默认返回最后一行的结果）：

FormatTMtCensusData <- function (tree, database.code) {

if (database.code == 'NOR') {
   census_info <- unique (tree[,c('database.code','tmt.census.id','census.id', 'census.date','census.n','tmt.plot.id','plot.id')])
   census_info$census.contact <- 'NA'
    census_info$census.contact.email <- 'NA'
    
 }
 census_info
}

同时建议处理database.code不为'NOR'的情况，避免返回NULL导致后续出错，比如可以添加默认返回值：

FormatTMtCensusData <- function (tree, database.code) {
  census_info <- NULL
  if (database.code == 'NOR') {
     census_info <- unique (tree[,c('database.code','tmt.census.id','census.id', 'census.date','census.n','tmt.plot.id','plot.id')])
     census_info$census.contact <- 'NA'
      census_info$census.contact.email <- 'NA'
      
   }
  census_info
}

内容的提问来源于stack exchange，提问作者starski