问题
Iterables present two methods for getLast
public static <T> T getLast(Iterable<T> iterable);
public static <T> T getLast(Iterable<T> iterable, @Nullable T defaultValue);
but only one for getFirst
public static <T> T getFirst(Iterable<T> iterable, @Nullable T defaultValue);
Is there are any design/implementation reason for breaking symmetry?
回答1:
I think the point is that there is no reason for a getFirst(iterable)
in that this could be done with iterable.iterator().next()
. Guava makes an excellent attempt to keep the API small and so does not add things that could / should be done easily another way.
On the other hand, there is not already a mechanism to test if an iterable is empty and if so return a default value instead of the first value. Hence, getFirst(iterable, default)
.
Also, there is not a simple way to get the last element, hence getLast(iterable)
and getLast(iterable, default)
回答2:
As an additions to @JohnB's answer I'd like to show Guava's devs opinion about getFirst(iterable). Kevin Bourrillion (head Guava's dev) writes there:
iterable.iterator().next() is perfectly clear and readable and unambiguous. I know exactly what it does, whereas with Iterators.getFirst(), I have to run off and look up how that library designer decided to do it.
Also, your notion of consistency is deeply misguided. We use consistency in how we present important functionality, but we never use it to justify adding worthless functionality, and you shouldn't in your own libraries either!
So, you have a choice:
- using
iterable.iterator().next()
, - using
Iterables.getFirst(Iterable<T> iterable, T default)
, - using
Iterables.get(Iterable<T>, 0)
, - writing your own method (probably containing
iterable.iterator().next()
and some docs) and use it as i.e.Iterables2.getFirst(iterable)
, - waiting for Kevin to change his mind ;)
PS: I had similar doubt some time ago and found exact duplicate of this question at that time.
来源:https://stackoverflow.com/questions/8473826/why-there-is-no-getfirstiterable-method