问题
Consider the following code:
IEnumerable<int> xx = null;
var tt = xx?.Where(x => x > 2).Select(x => x.ToString());
It assigns null
to tt
.
The question is: why does it work properly?
I thought I must use ?.
before Select as ?.Where(...)
returns null
.
Besides, if I split the second line into two separate lines:
IEnumerable<int> xx = null;
var yy = xx?.Where(x => x > 2);
var zz = yy.Select(x => x.ToString());
There will be the ArgumentNullException
on the third line as yy == null
.
What's the magic? :)
If this is because of short-circuiting, I've never thought that it can act like this.
回答1:
Yes, this is due to short-circuiting. From the MSDN reference:
...[T]he null-condition operators are short-circuiting. If one operation in a chain of conditional member access and index operation returns null, then the rest of the chain’s execution stops.
The reason your second example throws is because you have separate unchained statements. Short-circuiting cannot be applied across multiple statements.
回答2:
The null-conditional operator or also known as the null propagation operator is short-circuiting i.e if one operation in the chain:
var tt = xx?.Where(x => x > 2).Select(x => x.ToString());
returns null
, then the rest of the chain’s execution stops.
So in the above example Where
is never invoked as xx
is null
.
As for the second example, you're getting an ArgumentNullException
because that's the behaviour of extension methods. in this specific case, the Select
throws a ArgumentNullException
when the source or the provided selector is null
.
来源:https://stackoverflow.com/questions/48831588/why-can-i-omit-the-subsequent-null-conditional-operators-in-an-invocation-chain