Struct IndexSet

pub struct IndexSet<T, S = RandomState> { /* private fields */ }

A hash set where the iteration order of the values is independent of their hash values.

The interface is closely compatible with the standard HashSet, but also has additional features.

Order

The values have a consistent order that is determined by the sequence of insertion and removal calls on the set. The order does not depend on the values or the hash function at all. Note that insertion order and value are not affected if a re-insertion is attempted once an element is already present.

All iterators traverse the set in order. Set operation iterators like IndexSet::union produce a concatenated order, as do their matching “bitwise” operators. See their documentation for specifics.

The insertion order is preserved, with notable exceptions like the .remove() or .swap_remove() methods. Methods such as .sort_by() of course result in a new order, depending on the sorting order.

Indices

The values are indexed in a compact range without holes in the range 0..self.len(). For example, the method .get_full looks up the index for a value, and the method .get_index looks up the value by index.

Complexity

Internally, IndexSet<T, S> just holds an IndexMap<T, (), S>. Thus the complexity of the two are the same for most methods.

Examples

use indexmap::IndexSet;

// Collects which letters appear in a sentence.
let letters: IndexSet<_> = "a short treatise on fungi".chars().collect();

assert!(letters.contains(&'s'));
assert!(letters.contains(&'t'));
assert!(letters.contains(&'u'));
assert!(!letters.contains(&'y'));

Implementations

impl<T> IndexSet<T>

pub fn new() -> Self

Available on crate feature std only.

Create a new set. (Does not allocate.)

pub fn with_capacity(n: usize) -> Self

Available on crate feature std only.

Create a new set with capacity for n elements. (Does not allocate if n is zero.)

Computes in O(n) time.

impl<T, S> IndexSet<T, S>

pub fn with_capacity_and_hasher(n: usize, hash_builder: S) -> Self

Create a new set with capacity for n elements. (Does not allocate if n is zero.)

Computes in O(n) time.

pub const fn with_hasher(hash_builder: S) -> Self

Create a new set with hash_builder.

This function is const, so it can be called in static contexts.

pub fn capacity(&self) -> usize

Return the number of elements the set can hold without reallocating.

This number is a lower bound; the set might be able to hold more, but is guaranteed to be able to hold at least this many.

Computes in O(1) time.

pub fn hasher(&self) -> &S

Return a reference to the set’s BuildHasher.

pub fn len(&self) -> usize

Return the number of elements in the set.

Computes in O(1) time.

pub fn is_empty(&self) -> bool

Returns true if the set contains no elements.

Computes in O(1) time.

pub fn iter(&self) -> Iter<'_, T>

Return an iterator over the values of the set, in their order

pub fn clear(&mut self)

Remove all elements in the set, while preserving its capacity.

Computes in O(n) time.

pub fn truncate(&mut self, len: usize)

Shortens the set, keeping the first len elements and dropping the rest.

If len is greater than the set’s current length, this has no effect.

pub fn drain<R>(&mut self, range: R) -> Drain<'_, T>

where R: RangeBounds<usize>,

Clears the IndexSet in the given index range, returning those values as a drain iterator.

The range may be any type that implements RangeBounds<usize>, including all of the std::ops::Range* types, or even a tuple pair of Bound start and end values. To drain the set entirely, use RangeFull like set.drain(..).

This shifts down all entries following the drained range to fill the gap, and keeps the allocated memory for reuse.

Panics if the starting point is greater than the end point or if the end point is greater than the length of the set.

pub fn split_off(&mut self, at: usize) -> Self

where S: Clone,

Splits the collection into two at the given index.

Returns a newly allocated set containing the elements in the range [at, len). After the call, the original set will be left containing the elements [0, at) with its previous capacity unchanged.

Panics if at > len.

pub fn reserve(&mut self, additional: usize)

Reserve capacity for additional more values.

Computes in O(n) time.

pub fn reserve_exact(&mut self, additional: usize)

Reserve capacity for additional more values, without over-allocating.

Unlike reserve, this does not deliberately over-allocate the entry capacity to avoid frequent re-allocations. However, the underlying data structures may still have internal capacity requirements, and the allocator itself may give more space than requested, so this cannot be relied upon to be precisely minimal.

Computes in O(n) time.

pub fn try_reserve(&mut self, additional: usize) -> Result<(), TryReserveError>

Try to reserve capacity for additional more values.

Computes in O(n) time.

pub fn try_reserve_exact( &mut self, additional: usize, ) -> Result<(), TryReserveError>

Try to reserve capacity for additional more values, without over-allocating.

Unlike try_reserve, this does not deliberately over-allocate the entry capacity to avoid frequent re-allocations. However, the underlying data structures may still have internal capacity requirements, and the allocator itself may give more space than requested, so this cannot be relied upon to be precisely minimal.

Computes in O(n) time.

pub fn shrink_to_fit(&mut self)

Shrink the capacity of the set as much as possible.

Computes in O(n) time.

pub fn shrink_to(&mut self, min_capacity: usize)

Shrink the capacity of the set with a lower limit.

Computes in O(n) time.

impl<T, S> IndexSet<T, S>

where T: Hash + Eq, S: BuildHasher,

pub fn insert(&mut self, value: T) -> bool

Insert the value into the set.

If an equivalent item already exists in the set, it returns false leaving the original value in the set and without altering its insertion order. Otherwise, it inserts the new item and returns true.

Computes in O(1) time (amortized average).

pub fn insert_full(&mut self, value: T) -> (usize, bool)

Insert the value into the set, and get its index.

If an equivalent item already exists in the set, it returns the index of the existing item and false, leaving the original value in the set and without altering its insertion order. Otherwise, it inserts the new item and returns the index of the inserted item and true.

Computes in O(1) time (amortized average).

pub fn insert_sorted(&mut self, value: T) -> (usize, bool)

where T: Ord,

Insert the value into the set at its ordered position among sorted values.

This is equivalent to finding the position with binary_search, and if needed calling insert_before for a new value.

If the sorted item is found in the set, it returns the index of that existing item and false, without any change. Otherwise, it inserts the new item and returns its sorted index and true.

If the existing items are not already sorted, then the insertion index is unspecified (like slice::binary_search), but the value is moved to or inserted at that position regardless.

Computes in O(n) time (average). Instead of repeating calls to insert_sorted, it may be faster to call batched insert or extend and only call sort or sort_unstable once.

pub fn insert_before(&mut self, index: usize, value: T) -> (usize, bool)

Insert the value into the set before the value at the given index, or at the end.

If an equivalent item already exists in the set, it returns false leaving the original value in the set, but moved to the new position. The returned index will either be the given index or one less, depending on how the value moved. (See shift_insert for different behavior here.)

Otherwise, it inserts the new value exactly at the given index and returns true.

Panics if index is out of bounds. Valid indices are 0..=set.len() (inclusive).

Computes in O(n) time (average).

Examples

use indexmap::IndexSet;
let mut set: IndexSet<char> = ('a'..='z').collect();

// The new value '*' goes exactly at the given index.
assert_eq!(set.get_index_of(&'*'), None);
assert_eq!(set.insert_before(10, '*'), (10, true));
assert_eq!(set.get_index_of(&'*'), Some(10));

// Moving the value 'a' up will shift others down, so this moves *before* 10 to index 9.
assert_eq!(set.insert_before(10, 'a'), (9, false));
assert_eq!(set.get_index_of(&'a'), Some(9));
assert_eq!(set.get_index_of(&'*'), Some(10));

// Moving the value 'z' down will shift others up, so this moves to exactly 10.
assert_eq!(set.insert_before(10, 'z'), (10, false));
assert_eq!(set.get_index_of(&'z'), Some(10));
assert_eq!(set.get_index_of(&'*'), Some(11));

// Moving or inserting before the endpoint is also valid.
assert_eq!(set.len(), 27);
assert_eq!(set.insert_before(set.len(), '*'), (26, false));
assert_eq!(set.get_index_of(&'*'), Some(26));
assert_eq!(set.insert_before(set.len(), '+'), (27, true));
assert_eq!(set.get_index_of(&'+'), Some(27));
assert_eq!(set.len(), 28);

pub fn shift_insert(&mut self, index: usize, value: T) -> bool

Insert the value into the set at the given index.

If an equivalent item already exists in the set, it returns false leaving the original value in the set, but moved to the given index. Note that existing values cannot be moved to index == set.len()! (See insert_before for different behavior here.)

Otherwise, it inserts the new value at the given index and returns true.

Panics if index is out of bounds. Valid indices are 0..set.len() (exclusive) when moving an existing value, or 0..=set.len() (inclusive) when inserting a new value.

Computes in O(n) time (average).

Examples

use indexmap::IndexSet;
let mut set: IndexSet<char> = ('a'..='z').collect();

// The new value '*' goes exactly at the given index.
assert_eq!(set.get_index_of(&'*'), None);
assert_eq!(set.shift_insert(10, '*'), true);
assert_eq!(set.get_index_of(&'*'), Some(10));

// Moving the value 'a' up to 10 will shift others down, including the '*' that was at 10.
assert_eq!(set.shift_insert(10, 'a'), false);
assert_eq!(set.get_index_of(&'a'), Some(10));
assert_eq!(set.get_index_of(&'*'), Some(9));

// Moving the value 'z' down to 9 will shift others up, including the '*' that was at 9.
assert_eq!(set.shift_insert(9, 'z'), false);
assert_eq!(set.get_index_of(&'z'), Some(9));
assert_eq!(set.get_index_of(&'*'), Some(10));

// Existing values can move to len-1 at most, but new values can insert at the endpoint.
assert_eq!(set.len(), 27);
assert_eq!(set.shift_insert(set.len() - 1, '*'), false);
assert_eq!(set.get_index_of(&'*'), Some(26));
assert_eq!(set.shift_insert(set.len(), '+'), true);
assert_eq!(set.get_index_of(&'+'), Some(27));
assert_eq!(set.len(), 28);

use indexmap::IndexSet;
let mut set: IndexSet<char> = ('a'..='z').collect();

// This is an invalid index for moving an existing value!
set.shift_insert(set.len(), 'a');

pub fn replace(&mut self, value: T) -> Option<T>

Adds a value to the set, replacing the existing value, if any, that is equal to the given one, without altering its insertion order. Returns the replaced value.

Computes in O(1) time (average).

pub fn replace_full(&mut self, value: T) -> (usize, Option<T>)

Adds a value to the set, replacing the existing value, if any, that is equal to the given one, without altering its insertion order. Returns the index of the item and its replaced value.

Computes in O(1) time (average).

pub fn difference<'a, S2>( &'a self, other: &'a IndexSet<T, S2>, ) -> Difference<'a, T, S2>

where S2: BuildHasher,

Return an iterator over the values that are in self but not other.

Values are produced in the same order that they appear in self.

pub fn symmetric_difference<'a, S2>( &'a self, other: &'a IndexSet<T, S2>, ) -> SymmetricDifference<'a, T, S, S2>

where S2: BuildHasher,

Return an iterator over the values that are in self or other, but not in both.

Values from self are produced in their original order, followed by values from other in their original order.

pub fn intersection<'a, S2>( &'a self, other: &'a IndexSet<T, S2>, ) -> Intersection<'a, T, S2>

where S2: BuildHasher,

Return an iterator over the values that are in both self and other.

Values are produced in the same order that they appear in self.

pub fn union<'a, S2>(&'a self, other: &'a IndexSet<T, S2>) -> Union<'a, T, S>

where S2: BuildHasher,

Return an iterator over all values that are in self or other.

Values from self are produced in their original order, followed by values that are unique to other in their original order.

pub fn splice<R, I>( &mut self, range: R, replace_with: I, ) -> Splice<'_, I::IntoIter, T, S>

where R: RangeBounds<usize>, I: IntoIterator<Item = T>,

Creates a splicing iterator that replaces the specified range in the set with the given replace_with iterator and yields the removed items. replace_with does not need to be the same length as range.

The range is removed even if the iterator is not consumed until the end. It is unspecified how many elements are removed from the set if the Splice value is leaked.

The input iterator replace_with is only consumed when the Splice value is dropped. If a value from the iterator matches an existing entry in the set (outside of range), then the original will be unchanged. Otherwise, the new value will be inserted in the replaced range.

Panics if the starting point is greater than the end point or if the end point is greater than the length of the set.

Examples

use indexmap::IndexSet;

let mut set = IndexSet::from([0, 1, 2, 3, 4]);
let new = [5, 4, 3, 2, 1];
let removed: Vec<_> = set.splice(2..4, new).collect();

// 1 and 4 kept their positions, while 5, 3, and 2 were newly inserted.
assert!(set.into_iter().eq([0, 1, 5, 3, 2, 4]));
assert_eq!(removed, &[2, 3]);

pub fn append<S2>(&mut self, other: &mut IndexSet<T, S2>)

Moves all values from other into self, leaving other empty.

This is equivalent to calling insert for each value from other in order, which means that values that already exist in self are unchanged in their current position.

See also union to iterate the combined values by reference, without modifying self or other.

Examples

use indexmap::IndexSet;

let mut a = IndexSet::from([3, 2, 1]);
let mut b = IndexSet::from([3, 4, 5]);
let old_capacity = b.capacity();

a.append(&mut b);

assert_eq!(a.len(), 5);
assert_eq!(b.len(), 0);
assert_eq!(b.capacity(), old_capacity);

assert!(a.iter().eq(&[3, 2, 1, 4, 5]));

impl<T, S> IndexSet<T, S>

where S: BuildHasher,

pub fn contains<Q>(&self, value: &Q) -> bool

where Q: ?Sized + Hash + Equivalent<T>,

Return true if an equivalent to value exists in the set.

Computes in O(1) time (average).

pub fn get<Q>(&self, value: &Q) -> Option<&T>

where Q: ?Sized + Hash + Equivalent<T>,

Return a reference to the value stored in the set, if it is present, else None.

Computes in O(1) time (average).

pub fn get_full<Q>(&self, value: &Q) -> Option<(usize, &T)>

where Q: ?Sized + Hash + Equivalent<T>,

Return item index and value

pub fn get_index_of<Q>(&self, value: &Q) -> Option<usize>

where Q: ?Sized + Hash + Equivalent<T>,

Return item index, if it exists in the set

Computes in O(1) time (average).

pub fn remove<Q>(&mut self, value: &Q) -> bool

where Q: ?Sized + Hash + Equivalent<T>,

👎Deprecated: remove disrupts the set order – use swap_remove or shift_remove for explicit behavior.

Remove the value from the set, and return true if it was present.

NOTE: This is equivalent to .swap_remove(value), replacing this value’s position with the last element, and it is deprecated in favor of calling that explicitly. If you need to preserve the relative order of the values in the set, use .shift_remove(value) instead.

pub fn swap_remove<Q>(&mut self, value: &Q) -> bool

where Q: ?Sized + Hash + Equivalent<T>,

Remove the value from the set, and return true if it was present.

Like Vec::swap_remove, the value is removed by swapping it with the last element of the set and popping it off. This perturbs the position of what used to be the last element!

Return false if value was not in the set.

Computes in O(1) time (average).

pub fn shift_remove<Q>(&mut self, value: &Q) -> bool

where Q: ?Sized + Hash + Equivalent<T>,

Remove the value from the set, and return true if it was present.

Like Vec::remove, the value is removed by shifting all of the elements that follow it, preserving their relative order. This perturbs the index of all of those elements!

Return false if value was not in the set.

Computes in O(n) time (average).

pub fn take<Q>(&mut self, value: &Q) -> Option<T>

where Q: ?Sized + Hash + Equivalent<T>,

👎Deprecated: take disrupts the set order – use swap_take or shift_take for explicit behavior.

Removes and returns the value in the set, if any, that is equal to the given one.

NOTE: This is equivalent to .swap_take(value), replacing this value’s position with the last element, and it is deprecated in favor of calling that explicitly. If you need to preserve the relative order of the values in the set, use .shift_take(value) instead.

pub fn swap_take<Q>(&mut self, value: &Q) -> Option<T>

where Q: ?Sized + Hash + Equivalent<T>,

Removes and returns the value in the set, if any, that is equal to the given one.

Like Vec::swap_remove, the value is removed by swapping it with the last element of the set and popping it off. This perturbs the position of what used to be the last element!

Return None if value was not in the set.

Computes in O(1) time (average).

pub fn shift_take<Q>(&mut self, value: &Q) -> Option<T>

where Q: ?Sized + Hash + Equivalent<T>,

Removes and returns the value in the set, if any, that is equal to the given one.

Like Vec::remove, the value is removed by shifting all of the elements that follow it, preserving their relative order. This perturbs the index of all of those elements!

Return None if value was not in the set.

Computes in O(n) time (average).

pub fn swap_remove_full<Q>(&mut self, value: &Q) -> Option<(usize, T)>

where Q: ?Sized + Hash + Equivalent<T>,

Remove the value from the set return it and the index it had.

Like Vec::swap_remove, the value is removed by swapping it with the last element of the set and popping it off. This perturbs the position of what used to be the last element!

Return None if value was not in the set.

pub fn shift_remove_full<Q>(&mut self, value: &Q) -> Option<(usize, T)>

where Q: ?Sized + Hash + Equivalent<T>,

Remove the value from the set return it and the index it had.

Like Vec::remove, the value is removed by shifting all of the elements that follow it, preserving their relative order. This perturbs the index of all of those elements!

Return None if value was not in the set.

impl<T, S> IndexSet<T, S>

pub fn pop(&mut self) -> Option<T>

Remove the last value

This preserves the order of the remaining elements.

Computes in O(1) time (average).

pub fn retain<F>(&mut self, keep: F)

where F: FnMut(&T) -> bool,

Scan through each value in the set and keep those where the closure keep returns true.

The elements are visited in order, and remaining elements keep their order.

Computes in O(n) time (average).

pub fn sort(&mut self)

where T: Ord,

Sort the set’s values by their default ordering.

This is a stable sort – but equivalent values should not normally coexist in a set at all, so sort_unstable is preferred because it is generally faster and doesn’t allocate auxiliary memory.

See sort_by for details.

pub fn sort_by<F>(&mut self, cmp: F)

where F: FnMut(&T, &T) -> Ordering,

Sort the set’s values in place using the comparison function cmp.

Computes in O(n log n) time and O(n) space. The sort is stable.

pub fn sorted_by<F>(self, cmp: F) -> IntoIter<T>

where F: FnMut(&T, &T) -> Ordering,

Sort the values of the set and return a by-value iterator of the values with the result.

The sort is stable.

pub fn sort_unstable(&mut self)

where T: Ord,

Sort the set’s values by their default ordering.

See sort_unstable_by for details.

pub fn sort_unstable_by<F>(&mut self, cmp: F)

where F: FnMut(&T, &T) -> Ordering,

Sort the set’s values in place using the comparison function cmp.

Computes in O(n log n) time. The sort is unstable.

pub fn sorted_unstable_by<F>(self, cmp: F) -> IntoIter<T>

where F: FnMut(&T, &T) -> Ordering,

Sort the values of the set and return a by-value iterator of the values with the result.

pub fn sort_by_cached_key<K, F>(&mut self, sort_key: F)

where K: Ord, F: FnMut(&T) -> K,

Sort the set’s values in place using a key extraction function.

During sorting, the function is called at most once per entry, by using temporary storage to remember the results of its evaluation. The order of calls to the function is unspecified and may change between versions of indexmap or the standard library.

Computes in O(m n + n log n + c) time () and O(n) space, where the function is O(m), n is the length of the map, and c the capacity. The sort is stable.

pub fn binary_search(&self, x: &T) -> Result<usize, usize>

where T: Ord,

Search over a sorted set for a value.

Returns the position where that value is present, or the position where it can be inserted to maintain the sort. See slice::binary_search for more details.

Computes in O(log(n)) time, which is notably less scalable than looking the value up using get_index_of, but this can also position missing values.

pub fn binary_search_by<'a, F>(&'a self, f: F) -> Result<usize, usize>

where F: FnMut(&'a T) -> Ordering,

Search over a sorted set with a comparator function.

Returns the position where that value is present, or the position where it can be inserted to maintain the sort. See slice::binary_search_by for more details.

Computes in O(log(n)) time.

pub fn binary_search_by_key<'a, B, F>( &'a self, b: &B, f: F, ) -> Result<usize, usize>

where F: FnMut(&'a T) -> B, B: Ord,

Search over a sorted set with an extraction function.

Returns the position where that value is present, or the position where it can be inserted to maintain the sort. See slice::binary_search_by_key for more details.

Computes in O(log(n)) time.

pub fn partition_point<P>(&self, pred: P) -> usize

where P: FnMut(&T) -> bool,

Returns the index of the partition point of a sorted set according to the given predicate (the index of the first element of the second partition).

See slice::partition_point for more details.

Computes in O(log(n)) time.

pub fn reverse(&mut self)

Reverses the order of the set’s values in place.

Computes in O(n) time and O(1) space.

pub fn as_slice(&self) -> &Slice<T>

Returns a slice of all the values in the set.

Computes in O(1) time.

pub fn into_boxed_slice(self) -> Box<Slice<T>>

Converts into a boxed slice of all the values in the set.

Note that this will drop the inner hash table and any excess capacity.

pub fn get_index(&self, index: usize) -> Option<&T>

Get a value by index

Valid indices are 0 <= index < self.len().

Computes in O(1) time.

pub fn get_range<R: RangeBounds<usize>>(&self, range: R) -> Option<&Slice<T>>

Returns a slice of values in the given range of indices.

Valid indices are 0 <= index < self.len().

Computes in O(1) time.

pub fn first(&self) -> Option<&T>

Get the first value

Computes in O(1) time.

pub fn last(&self) -> Option<&T>

Get the last value

Computes in O(1) time.

pub fn swap_remove_index(&mut self, index: usize) -> Option<T>

Remove the value by index

Valid indices are 0 <= index < self.len().

Like Vec::swap_remove, the value is removed by swapping it with the last element of the set and popping it off. This perturbs the position of what used to be the last element!

Computes in O(1) time (average).

pub fn shift_remove_index(&mut self, index: usize) -> Option<T>

Remove the value by index

Valid indices are 0 <= index < self.len().

Like Vec::remove, the value is removed by shifting all of the elements that follow it, preserving their relative order. This perturbs the index of all of those elements!

Computes in O(n) time (average).

pub fn move_index(&mut self, from: usize, to: usize)

Moves the position of a value from one index to another by shifting all other values in-between.

• If from < to, the other values will shift down while the targeted value moves up.
• If from > to, the other values will shift up while the targeted value moves down.

Panics if from or to are out of bounds.

Computes in O(n) time (average).

pub fn swap_indices(&mut self, a: usize, b: usize)

Swaps the position of two values in the set.

Panics if a or b are out of bounds.

Computes in O(1) time (average).

impl<T, S> IndexSet<T, S>

where T: Eq + Hash, S: BuildHasher,

pub fn is_disjoint<S2>(&self, other: &IndexSet<T, S2>) -> bool

where S2: BuildHasher,

Returns true if self has no elements in common with other.

pub fn is_subset<S2>(&self, other: &IndexSet<T, S2>) -> bool

where S2: BuildHasher,

Returns true if all elements of self are contained in other.

pub fn is_superset<S2>(&self, other: &IndexSet<T, S2>) -> bool

where S2: BuildHasher,

Returns true if all elements of other are contained in self.

impl<T, S> IndexSet<T, S>

where T: Hash + Eq + Sync, S: BuildHasher + Sync,

Parallel iterator methods and other parallel methods.

The following methods require crate feature "rayon".

See also the IntoParallelIterator implementations.

pub fn par_difference<'a, S2>( &'a self, other: &'a IndexSet<T, S2>, ) -> ParDifference<'a, T, S, S2>

where S2: BuildHasher + Sync,

Available on crate feature rayon only.

Return a parallel iterator over the values that are in self but not other.

While parallel iterators can process items in any order, their relative order in the self set is still preserved for operations like reduce and collect.

pub fn par_symmetric_difference<'a, S2>( &'a self, other: &'a IndexSet<T, S2>, ) -> ParSymmetricDifference<'a, T, S, S2>

where S2: BuildHasher + Sync,

Available on crate feature rayon only.

Return a parallel iterator over the values that are in self or other, but not in both.

While parallel iterators can process items in any order, their relative order in the sets is still preserved for operations like reduce and collect. Values from self are produced in their original order, followed by values from other in their original order.

pub fn par_intersection<'a, S2>( &'a self, other: &'a IndexSet<T, S2>, ) -> ParIntersection<'a, T, S, S2>

where S2: BuildHasher + Sync,

Available on crate feature rayon only.

Return a parallel iterator over the values that are in both self and other.

While parallel iterators can process items in any order, their relative order in the self set is still preserved for operations like reduce and collect.

pub fn par_union<'a, S2>( &'a self, other: &'a IndexSet<T, S2>, ) -> ParUnion<'a, T, S, S2>

where S2: BuildHasher + Sync,

Available on crate feature rayon only.

Return a parallel iterator over all values that are in self or other.

While parallel iterators can process items in any order, their relative order in the sets is still preserved for operations like reduce and collect. Values from self are produced in their original order, followed by values that are unique to other in their original order.

pub fn par_eq<S2>(&self, other: &IndexSet<T, S2>) -> bool

where S2: BuildHasher + Sync,

Available on crate feature rayon only.

Returns true if self contains all of the same values as other, regardless of each set’s indexed order, determined in parallel.

pub fn par_is_disjoint<S2>(&self, other: &IndexSet<T, S2>) -> bool

where S2: BuildHasher + Sync,

Available on crate feature rayon only.

Returns true if self has no elements in common with other, determined in parallel.

pub fn par_is_superset<S2>(&self, other: &IndexSet<T, S2>) -> bool

where S2: BuildHasher + Sync,

Available on crate feature rayon only.

Returns true if all elements of other are contained in self, determined in parallel.

pub fn par_is_subset<S2>(&self, other: &IndexSet<T, S2>) -> bool

where S2: BuildHasher + Sync,

Available on crate feature rayon only.

Returns true if all elements of self are contained in other, determined in parallel.

impl<T, S> IndexSet<T, S>

where T: Send,

Parallel sorting methods.

The following methods require crate feature "rayon".

pub fn par_sort(&mut self)

where T: Ord,

Available on crate feature rayon only.

Sort the set’s values in parallel by their default ordering.

pub fn par_sort_by<F>(&mut self, cmp: F)

where F: Fn(&T, &T) -> Ordering + Sync,

Available on crate feature rayon only.

Sort the set’s values in place and in parallel, using the comparison function cmp.

pub fn par_sorted_by<F>(self, cmp: F) -> IntoParIter<T>

where F: Fn(&T, &T) -> Ordering + Sync,

Available on crate feature rayon only.

Sort the values of the set in parallel and return a by-value parallel iterator of the values with the result.

pub fn par_sort_unstable(&mut self)

where T: Ord,

Available on crate feature rayon only.

Sort the set’s values in parallel by their default ordering.

pub fn par_sort_unstable_by<F>(&mut self, cmp: F)

where F: Fn(&T, &T) -> Ordering + Sync,

Available on crate feature rayon only.

Sort the set’s values in place and in parallel, using the comparison function cmp.

pub fn par_sorted_unstable_by<F>(self, cmp: F) -> IntoParIter<T>

where F: Fn(&T, &T) -> Ordering + Sync,

Available on crate feature rayon only.

Sort the values of the set in parallel and return a by-value parallel iterator of the values with the result.

pub fn par_sort_by_cached_key<K, F>(&mut self, sort_key: F)

where K: Ord + Send, F: Fn(&T) -> K + Sync,

Available on crate feature rayon only.

Sort the set’s values in place and in parallel, using a key extraction function.

Trait Implementations

impl<'a, T, S> Arbitrary<'a> for IndexSet<T, S>

where T: Arbitrary<'a> + Hash + Eq, S: BuildHasher + Default,

Available on crate feature arbitrary only.

fn arbitrary(u: &mut Unstructured<'a>) -> Result<Self>

Generate an arbitrary value of Self from the given unstructured data.

fn arbitrary_take_rest(u: Unstructured<'a>) -> Result<Self>

Generate an arbitrary value of Self from the entirety of the given unstructured data.

fn size_hint(depth: usize) -> (usize, Option<usize>)

Get a size hint for how many bytes out of an Unstructured this type needs to construct itself.

fn try_size_hint( depth: usize, ) -> Result<(usize, Option<usize>), MaxRecursionReached>

Get a size hint for how many bytes out of an Unstructured this type needs to construct itself.

impl<T, S> Arbitrary for IndexSet<T, S>

where T: Arbitrary + Hash + Eq, S: BuildHasher + Default + Clone + 'static,

Available on crate feature quickcheck only.

fn arbitrary(g: &mut Gen) -> Self

Return an arbitrary value.

fn shrink(&self) -> Box<dyn Iterator<Item = Self>>

Return an iterator of values that are smaller than itself.

impl<T, S1, S2> BitAnd<&IndexSet<T, S2>> for &IndexSet<T, S1>

where T: Eq + Hash + Clone, S1: BuildHasher + Default, S2: BuildHasher,

fn bitand(self, other: &IndexSet<T, S2>) -> Self::Output

Returns the set intersection, cloned into a new set.

Values are collected in the same order that they appear in self.

type Output = IndexSet<T, S1>

The resulting type after applying the & operator.

impl<T, S1, S2> BitOr<&IndexSet<T, S2>> for &IndexSet<T, S1>

where T: Eq + Hash + Clone, S1: BuildHasher + Default, S2: BuildHasher,

fn bitor(self, other: &IndexSet<T, S2>) -> Self::Output

Returns the set union, cloned into a new set.

Values from self are collected in their original order, followed by values that are unique to other in their original order.

type Output = IndexSet<T, S1>

The resulting type after applying the | operator.

impl<T, S1, S2> BitXor<&IndexSet<T, S2>> for &IndexSet<T, S1>

where T: Eq + Hash + Clone, S1: BuildHasher + Default, S2: BuildHasher,

fn bitxor(self, other: &IndexSet<T, S2>) -> Self::Output

Returns the set symmetric-difference, cloned into a new set.

Values from self are collected in their original order, followed by values from other in their original order.

type Output = IndexSet<T, S1>

The resulting type after applying the ^ operator.

impl<T, S> BorshDeserialize for IndexSet<T, S>

where T: BorshDeserialize + Eq + Hash, S: BuildHasher + Default,

Available on crate feature borsh only.

fn deserialize_reader<R: Read>(reader: &mut R) -> Result<Self>

fn deserialize(buf: &mut &[u8]) -> Result<Self, Error>

Deserializes this instance from a given slice of bytes. Updates the buffer to point at the remaining bytes.

fn try_from_slice(v: &[u8]) -> Result<Self, Error>

Deserialize this instance from a slice of bytes.

fn try_from_reader<R>(reader: &mut R) -> Result<Self, Error>

where R: Read,

impl<T, S> BorshSerialize for IndexSet<T, S>

where T: BorshSerialize,

Available on crate feature borsh only.

fn serialize<W: Write>(&self, writer: &mut W) -> Result<()>

impl<T, S> Clone for IndexSet<T, S>

where T: Clone, S: Clone,

fn clone(&self) -> Self

Returns a copy of the value.

fn clone_from(&mut self, other: &Self)

Performs copy-assignment from source.

impl<T, S> Debug for IndexSet<T, S>

where T: Debug,

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl<T, S> Default for IndexSet<T, S>

where S: Default,

fn default() -> Self

Return an empty IndexSet

impl<'de, T, S> Deserialize<'de> for IndexSet<T, S>

where T: Deserialize<'de> + Eq + Hash, S: Default + BuildHasher,

Available on crate feature serde only.

fn deserialize<D>(deserializer: D) -> Result<Self, D::Error>

where D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl<'a, T, S> Extend<&'a T> for IndexSet<T, S>

where T: Hash + Eq + Copy + 'a, S: BuildHasher,

fn extend<I: IntoIterator<Item = &'a T>>(&mut self, iterable: I)

Extends a collection with the contents of an iterator.

fn extend_one(&mut self, item: A)

🔬This is a nightly-only experimental API. (extend_one)

Extends a collection with exactly one element.

fn extend_reserve(&mut self, additional: usize)

🔬This is a nightly-only experimental API. (extend_one)

Reserves capacity in a collection for the given number of additional elements.

impl<T, S> Extend<T> for IndexSet<T, S>

where T: Hash + Eq, S: BuildHasher,

fn extend<I: IntoIterator<Item = T>>(&mut self, iterable: I)

Extends a collection with the contents of an iterator.

fn extend_one(&mut self, item: A)

🔬This is a nightly-only experimental API. (extend_one)

Extends a collection with exactly one element.

fn extend_reserve(&mut self, additional: usize)

🔬This is a nightly-only experimental API. (extend_one)

Reserves capacity in a collection for the given number of additional elements.

impl<T, const N: usize> From<[T; N]> for IndexSet<T, RandomState>

where T: Eq + Hash,

Available on crate feature std only.

fn from(arr: [T; N]) -> Self

Examples

use indexmap::IndexSet;

let set1 = IndexSet::from([1, 2, 3, 4]);
let set2: IndexSet<_> = [1, 2, 3, 4].into();
assert_eq!(set1, set2);

impl<T, S> FromIterator<T> for IndexSet<T, S>

where T: Hash + Eq, S: BuildHasher + Default,

fn from_iter<I: IntoIterator<Item = T>>(iterable: I) -> Self

Creates a value from an iterator.

impl<T, S> FromParallelIterator<T> for IndexSet<T, S>

where T: Eq + Hash + Send, S: BuildHasher + Default + Send,

Available on crate feature rayon only.

fn from_par_iter<I>(iter: I) -> Self

where I: IntoParallelIterator<Item = T>,

Creates an instance of the collection from the parallel iterator par_iter.

impl<T, S> Index<(Bound<usize>, Bound<usize>)> for IndexSet<T, S>

type Output = Slice<T>

The returned type after indexing.

fn index(&self, range: (Bound<usize>, Bound<usize>)) -> &Self::Output

Performs the indexing (container[index]) operation.

impl<T, S> Index<Range<usize>> for IndexSet<T, S>

type Output = Slice<T>

The returned type after indexing.

fn index(&self, range: Range<usize>) -> &Self::Output

Performs the indexing (container[index]) operation.

impl<T, S> Index<RangeFrom<usize>> for IndexSet<T, S>

type Output = Slice<T>

The returned type after indexing.

fn index(&self, range: RangeFrom<usize>) -> &Self::Output

Performs the indexing (container[index]) operation.

impl<T, S> Index<RangeFull> for IndexSet<T, S>

type Output = Slice<T>

The returned type after indexing.

fn index(&self, range: RangeFull) -> &Self::Output

Performs the indexing (container[index]) operation.

impl<T, S> Index<RangeInclusive<usize>> for IndexSet<T, S>

type Output = Slice<T>

The returned type after indexing.

fn index(&self, range: RangeInclusive<usize>) -> &Self::Output

Performs the indexing (container[index]) operation.

impl<T, S> Index<RangeTo<usize>> for IndexSet<T, S>

type Output = Slice<T>

The returned type after indexing.

fn index(&self, range: RangeTo<usize>) -> &Self::Output

Performs the indexing (container[index]) operation.

impl<T, S> Index<RangeToInclusive<usize>> for IndexSet<T, S>

type Output = Slice<T>

The returned type after indexing.

fn index(&self, range: RangeToInclusive<usize>) -> &Self::Output

Performs the indexing (container[index]) operation.

impl<T, S> Index<usize> for IndexSet<T, S>

Access IndexSet values at indexed positions.

Examples

use indexmap::IndexSet;

let mut set = IndexSet::new();
for word in "Lorem ipsum dolor sit amet".split_whitespace() {
    set.insert(word.to_string());
}
assert_eq!(set[0], "Lorem");
assert_eq!(set[1], "ipsum");
set.reverse();
assert_eq!(set[0], "amet");
assert_eq!(set[1], "sit");
set.sort();
assert_eq!(set[0], "Lorem");
assert_eq!(set[1], "amet");

use indexmap::IndexSet;

let mut set = IndexSet::new();
set.insert("foo");
println!("{:?}", set[10]); // panics!

fn index(&self, index: usize) -> &T

Returns a reference to the value at the supplied index.

Panics if index is out of bounds.

type Output = T

The returned type after indexing.

impl<'de, T, S, E> IntoDeserializer<'de, E> for IndexSet<T, S>

where T: IntoDeserializer<'de, E> + Eq + Hash, S: BuildHasher, E: Error,

Available on crate feature serde only.

type Deserializer = SeqDeserializer<<IndexSet<T, S> as IntoIterator>::IntoIter, E>

The type of the deserializer being converted into.

fn into_deserializer(self) -> Self::Deserializer

Convert this value into a deserializer.

impl<'a, T, S> IntoIterator for &'a IndexSet<T, S>

type Item = &'a T

The type of the elements being iterated over.

type IntoIter = Iter<'a, T>

Which kind of iterator are we turning this into?

fn into_iter(self) -> Self::IntoIter

Creates an iterator from a value.

impl<T, S> IntoIterator for IndexSet<T, S>

type Item = T

The type of the elements being iterated over.

type IntoIter = IntoIter<T>

Which kind of iterator are we turning this into?

fn into_iter(self) -> Self::IntoIter

Creates an iterator from a value.

impl<'a, T, S> IntoParallelIterator for &'a IndexSet<T, S>

where T: Sync,

Available on crate feature rayon only.

type Item = &'a T

The type of item that the parallel iterator will produce.

type Iter = ParIter<'a, T>

The parallel iterator type that will be created.

fn into_par_iter(self) -> Self::Iter

Converts self into a parallel iterator.

impl<T, S> IntoParallelIterator for IndexSet<T, S>

where T: Send,

Available on crate feature rayon only.

type Item = T

The type of item that the parallel iterator will produce.

type Iter = IntoParIter<T>

The parallel iterator type that will be created.

fn into_par_iter(self) -> Self::Iter

Converts self into a parallel iterator.

impl<T, S> MutableValues for IndexSet<T, S>

where S: BuildHasher,

Opt-in mutable access to IndexSet values.

See MutableValues for more information.

type Value = T

fn get_full_mut2<Q>(&mut self, value: &Q) -> Option<(usize, &mut T)>

where Q: ?Sized + Hash + Equivalent<T>,

Return item index and mutable reference to the value

fn get_index_mut2(&mut self, index: usize) -> Option<&mut T>

Return mutable reference to the value at an index.

fn retain2<F>(&mut self, keep: F)

where F: FnMut(&mut T) -> bool,

Scan through each value in the set and keep those where the closure keep returns true.

impl<'a, T, S> ParallelDrainRange for &'a mut IndexSet<T, S>

where T: Send,

Available on crate feature rayon only.

type Item = T

The type of item that the parallel iterator will produce. This is usually the same as IntoParallelIterator::Item.

type Iter = ParDrain<'a, T>

The draining parallel iterator type that will be created.

fn par_drain<R: RangeBounds<usize>>(self, range: R) -> Self::Iter

Returns a draining parallel iterator over a range of the collection.

impl<'a, T, S> ParallelExtend<&'a T> for IndexSet<T, S>

where T: Copy + Eq + Hash + Send + Sync + 'a, S: BuildHasher + Send,

Available on crate feature rayon only.

fn par_extend<I>(&mut self, iter: I)

where I: IntoParallelIterator<Item = &'a T>,

Extends an instance of the collection with the elements drawn from the parallel iterator par_iter.

impl<T, S> ParallelExtend<T> for IndexSet<T, S>

where T: Eq + Hash + Send, S: BuildHasher + Send,

Available on crate feature rayon only.

fn par_extend<I>(&mut self, iter: I)

where I: IntoParallelIterator<Item = T>,

Extends an instance of the collection with the elements drawn from the parallel iterator par_iter.

impl<T, S1, S2> PartialEq<IndexSet<T, S2>> for IndexSet<T, S1>

where T: Hash + Eq, S1: BuildHasher, S2: BuildHasher,

fn eq(&self, other: &IndexSet<T, S2>) -> bool

Tests for self and other values to be equal, and is used by ==.

fn ne(&self, other: &Rhs) -> bool

Tests for !=. The default implementation is almost always sufficient, and should not be overridden without very good reason.

impl<T, S> Serialize for IndexSet<T, S>

where T: Serialize,

Available on crate feature serde only.

fn serialize<Se>(&self, serializer: Se) -> Result<Se::Ok, Se::Error>

where Se: Serializer,

Serialize this value into the given Serde serializer.

impl<T, S1, S2> Sub<&IndexSet<T, S2>> for &IndexSet<T, S1>

where T: Eq + Hash + Clone, S1: BuildHasher + Default, S2: BuildHasher,

fn sub(self, other: &IndexSet<T, S2>) -> Self::Output

Returns the set difference, cloned into a new set.

Values are collected in the same order that they appear in self.

type Output = IndexSet<T, S1>

The resulting type after applying the - operator.

impl<T, S> Eq for IndexSet<T, S>

where T: Eq + Hash, S: BuildHasher,