是否已记录R data.table通过引用作为参数传递？ [英] Is R data.table documented to pass by reference as argument?

问题描述

检查此玩具代码：

> x <- data.table(a = 1:2) 
> foo <- function(z) { z[, b:=3:4]  }
> y <- foo(x)
> x[]
   a b
1: 1 3
2: 2 4

似乎data.table通过引用传递。这是故意的吗？有记录吗？我确实通读了文档，却找不到关于此行为的提及。

It seems data.table is passed by reference. Is this intentional? Is this documented? I did read through the docs and couldn't find a mention of this behaviour.

我不是不是，它询问R的文档参考语义（在<$ c中$ c>：= ， set *** 等）。我问一个data.table完整对象是否应该作为函数参数通过引用传递。

I'm not asking about R's documented reference semantics (in :=, set*** and some others). I'm asking whether a data.table complete object is supposed to be passed by reference as a function argument.

编辑：按照@Oliver的回答，在这里是一些更奇怪的例子。

Following @Oliver's answer, here are some more curious examples.

> dt<- data.table(a=1:2)
> attr(dt, ".internal.selfref")
<pointer: 0x564776a93e88>
> address(dt)
[1] "0x5647bc0f6c50"
> 
> ff<-function(x) { x[, b:=3:4]; print(address(x)); print(attr(dt, ".internal.selfref")) }
> ff(dt)
[1] "0x5647bc0f6c50"
<pointer: 0x564776a93e88>

所以 .internal.selfref 不仅与呼叫者的相同dt复制，地址也是。确实是同一对象。 （我认为）。

So not only is .internal.selfref identical to the caller's dt copy, so is the address. It really is the same object. (I think).

对于data.frames并非完全如此：

This is not exactly the case for data.frames:

> df<- data.frame(a=1:2)
> address(df)
[1] "0x5647b39d21e8"
> ff<-function(x) { print(address(x)); x$b=3:4; print(address(x)) }
> 
> ff(df)
[1] "0x5647b39d21e8"
[1] "0x5647ae24de78"

也许根本问题是常规的data.table操作某种程度上不会触发R的修改时复制语义？ p>

Maybe the root issue is that regular data.table operations somehow do not trigger R's copy-on-modify semantics?

推荐答案

我认为您感到惊讶的实际上是R行为，这就是为什么在 data.table （也许应该仍然如此，因为对于 data.table 而言，影响更为重要）。

I think what you're being surprised about is actually R behavior, which is why it's not specifically documented in data.table (maybe it should be anyway, as the implications are more important for data.table).

让您惊讶的是，传递给函数的对象具有相同的地址，但是对于 base R也是如此：

You were surprised that the object passed to a function had the same address, but this is the same for base R as well:

x = 1:10
address(x)
# [1] "0x7fb7d4b6c820"
(function(y) {print(address(y))})(x)
# [1] "0x7fb7d4b6c820"

函数环境中已复制的是指针 x 。此外，对于 base R，父 x 是不可变的：

What's being copied in the function environment is the pointer to x. Moreover, for base R, the parent x is immutable:

foo = function(y) {
  print(address(y))
  y[1L] = 2L
  print(address(y))
}
foo(x)
# [1] "0x7fb7d4b6c820"
# [1] "0x7fb7d4e11d28"

也就是说，只要我们尝试编辑 y ，将进行复制。这与引用计数有关-您可以在此例如此演示文稿

That is, as soon as we try to edit y, a copy is made. This is related to reference counting -- you can see some work by Luke Tierney on this, e.g. this presentation

data.table 的区别是 data.table 为父对象启用编辑权限-我认为您知道这是一把双刃剑。

The difference for data.table is that data.table enables edit permissions for the parent object -- a double-edged sword as I think you know.

这篇关于是否已记录R data.table通过引用作为参数传递？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！