.unique() and .unique_by(..) operations #37
Conversation
Functions make iterators that return unique elements, storing visited elements in a hash set. `.unique_by(..)` uses supplied function to make keys, and `.unique()` uses `.clone()`.
|
Updated PR with explicit test. |
|
Thanks! You're taking over the task from PR #30, I guess. I'll leave some comments inline. Style wording, etc can all be fixed by me later if you don't want to, but incorrect size hint, that won't pass :-) |
|
|
||
| #[inline] | ||
| fn size_hint(&self) -> (usize, Option<usize>) { | ||
| self.iter.size_hint() |
There was a problem hiding this comment.
This should use the upper bound of the adapted iterator, and use a minimum of 0. Unless the hashset is empty, then use either 0 or 1 as lower bound depending on the adapted itertator.
Simply put -- our full information about how many elements remain, keeping in mind that they may all be duplicates or all unique.
There was a problem hiding this comment.
Thanks, I made a mistake.
I think 0 is fine for lower bound. I think in 99% cases size_hint is used on fresh new iterator, so hash-set would be empty.
|
Changed PR according to your comments. I didn't see #30 before you pointed to it. It's funny I made the same request independently. |
|
You managed to hit the requirements very well without having seen that PR then! Thanks for the contribution |
|
A possible follow up might be: A method to extract the hash set. You might as well allow it to be used after it was constructed? Not sure. |
.unique() and .unique_by(..) operations
|
Released in 0.3.16. |
|
Marvellous! Thanks for doing this :) |
Functions make iterators that return unique elements, storing visited
elements in a hash set.
.unique_by(..)uses supplied function to make keys, and.unique()uses
.clone().