为什么使用循环变量时地址会发生变化? [英] Why is the address of a loop variable changing when using it?

问题描述

程序1:

 library(pryr)

 for (x in 1:3) {
     print(c(address(x), refs(x)))
 }

输出例如:

[1] "0x7a6a6c8" "1"        
[1] "0x7a6a6c8" "1"        
[1] "0x7a6a6c8" "1"

程序2:

library(pryr)

for (x in 1:3) {
  print(c(address(x), refs(x)))
  print(x)
}

输出例如:

[1] "0x7ae0298" "1"        
[1] 1
[1] "0x7ae88c8" "1"        
[1] 2
[1] "0x7af2668" "1"        
[1] 3

显然，在程序2中x的值正在更改，但是为什么地址也更改?当for循环运行约500,000,000次而在循环期间未调用gc时，这是否会导致内存泄漏?

Obviously the value of x is changing in Program 2, but why is the adress changing too? Can this lead to memory leaks when having a for loop running about 500,000,000 times while the gc is not called during the loop?

推荐答案

在循环末尾具有print(x)会将其标记为多引用，如@alexis_laz所述.由于R是一种动态语言，因此这很容易发生.为了测试其效果，我们可以打印refs(x)，print(x)，refs(x)的输出:

Having print(x) at the end of the loop marks it as multi-reference, as @alexis_laz mentioned. Since R is a dynamic language, this can happen easily. To test the effects of this, we can print the output of refs(x), print(x), refs(x):

for (x in 1:3) {
  print(refs(x))
  print(x)
  print(refs(x)); cat("\n")
}

输出:

[1] 1
[1] 1
[1] 2

[1] 1
[1] 2
[1] 2

[1] 1
[1] 3
[1] 2

这篇关于为什么使用循环变量时地址会发生变化?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！