在函数内部但不在函数外部使用“alm()”时出现“找不到对象”错误

Question

我正在使用

alm()

包中的

greybox

函数来拟合位置和比例模型。但是，每当我将变量传递给

formula = 

的

alm()

参数时，我都会收到错误。这仅发生在函数内部。外部函数可以工作。返回的错误是“as.formula(form)中的错误：找不到对象'form'”。下面是一个可重现的示例。有修复的想法吗？ 可重现的示例：

library(greybox)
df <- data.frame(x = rnorm(100),
                 y = rnorm(100))

test_string = "y ~ x"

alm(as.formula(test_string), data = df)
#Passing formula to alm() works!

test_fun = function(form, df){
  #alm(scaled_meaning ~ short_dis_km, data = x)
  alm(as.formula(form), data = df)
}

test_fun(form = test_string, df = df)
#Passing formula to alm() does not work!

Answer 1

这可行，但这是一项肮脏的工作。 我认为这是一个与全球环境或其他相关的问题。如果将 form 声明为全局变量，则可以避免此问题。但不干净。

library(greybox)
df <- data.frame(x = rnorm(100),
                 y = rnorm(100))

test_string = "y ~ x"

alm(as.formula(test_string), data = df)
#Passing formula to alm() works!

test_fun = function(form, df){
  #alm(scaled_meaning ~ short_dis_km, data = x)
  form <<- form
  f<-alm(as.formula(form), data = df)
  rm(form)
  f
}

test_fun(form = test_string, df = df)
#Passing formula to alm() does not work!

Answer 2

在包修复范围错误之前，您可以包装在 do.call 中，由于某种原因，这修复了范围问题：

library(greybox)
df <- data.frame(x = rnorm(100),
                 y = rnorm(100))

test_string = "y ~ x"

alm(as.formula(test_string), data = df)
#Passing formula to alm() works!

test_fun = function(form, df){
  #alm(scaled_meaning ~ short_dis_km, data = x)
  alm(as.formula(form), data = df)
}

test_fun(form = test_string, df = df)
#Passing formula to alm() does not work!

test_fun2 = function(form, df){
  #alm(scaled_meaning ~ short_dis_km, data = x)
  do.call(alm,list(as.formula(form),data=df))
}

test_fun2(form = test_string, df=df)
# This one works!