`ifelse` 和 Julia 中的三元运算符有什么区别? [英] What is the difference between `ifelse` and the ternary operator in Julia?

问题描述

假设我有这个代码:

cond = true
a = cond ? 1 : 2
b = ifelse(cond, 1, 2)

这两种操作有什么区别?

What is the difference between the two operations?

推荐答案

在你写的例子中，没有有效的区别.但是，如果这两个分支比简单的整数文字更复杂，那么就会有所不同:

In the example you wrote, there is no effective difference. If, however, the two branches were more complicated than simple integer literals, then there is a difference:

julia> f() = (println("calling f()!"); 1)
f (generic function with 1 method)

julia> g() = (println("calling g()!"); 2)
g (generic function with 1 method)

julia> cond ? f() : g()
calling f()!
1

julia> ifelse(cond, f(), g())
calling f()!
calling g()!
1

换句话说，ifelse 只是一个普通的函数.就像所有其他函数一样，它的参数总是被评估.三元运算符的语法等价于:

In other words, ifelse is just a normal function. And just like all other functions, its arguments are always evaluated. The ternary operator is syntax equivalent to:

if cond
    f()
else
    g()
end

请注意，在某些情况下，这可能会导致处理器使用的指令不同(即，将分支更改为查找)，这可能会对两个分支中的代码成本(或如果在 @simd 循环中，则不是那么微妙)...但通常 Julia 和 LLVM 通常足够聪明，如果可能的话，无论哪种方式都可以做到最好.

Note that in some cases, this can result in a difference in the instructions your processor uses (that is, changing a branch to a lookup) which can have subtle performance implications beyond the costs of the code in the two branches (or not so subtle if inside a @simd loop)... but often Julia and LLVM are often smart enough to do the best thing either way if it's possible.

这篇关于`ifelse` 和 Julia 中的三元运算符有什么区别?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！