1 // Copyright 2014 The Rust Project Developers. See the COPYRIGHT
2 // file at the top-level directory of this distribution and at
3 // http://rust-lang.org/COPYRIGHT.
5 // Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or
6 // http://www.apache.org/licenses/LICENSE-2.0> or the MIT license
7 // <LICENSE-MIT or http://opensource.org/licenses/MIT>, at your
8 // option. This file may not be copied, modified, or distributed
9 // except according to those terms.
11 // This implementation is largely based on the high-level description and analysis of B-Trees
12 // found in *Open Data Structures* (ODS). Although our implementation does not use any of
13 // the source found in ODS, if one wishes to review the high-level design of this structure, it
14 // can be freely downloaded at http://opendatastructures.org/. Its contents are as of this
15 // writing (August 2014) freely licensed under the following Creative Commons Attribution
16 // License: [CC BY 2.5 CA](http://creativecommons.org/licenses/by/2.5/ca/).
18 pub use self::Entry::*;
22 use core::borrow::BorrowFrom;
23 use core::cmp::Ordering;
24 use core::default::Default;
26 use core::hash::{Writer, Hash};
27 use core::iter::{Map, FromIterator};
28 use core::ops::{Index, IndexMut};
29 use core::{iter, fmt, mem};
31 use ring_buf::RingBuf;
33 use self::Continuation::{Continue, Finished};
35 use super::node::ForceResult::{Leaf, Internal};
36 use super::node::TraversalItem::{self, Elem, Edge};
37 use super::node::{Traversal, MutTraversal, MoveTraversal};
38 use super::node::{self, Node, Found, GoDown};
40 // FIXME(conventions): implement bounded iterators
42 /// A map based on a B-Tree.
44 /// B-Trees represent a fundamental compromise between cache-efficiency and actually minimizing
45 /// the amount of work performed in a search. In theory, a binary search tree (BST) is the optimal
46 /// choice for a sorted map, as a perfectly balanced BST performs the theoretical minimum amount of
47 /// comparisons necessary to find an element (log<sub>2</sub>n). However, in practice the way this
48 /// is done is *very* inefficient for modern computer architectures. In particular, every element
49 /// is stored in its own individually heap-allocated node. This means that every single insertion
50 /// triggers a heap-allocation, and every single comparison should be a cache-miss. Since these
51 /// are both notably expensive things to do in practice, we are forced to at very least reconsider
54 /// A B-Tree instead makes each node contain B-1 to 2B-1 elements in a contiguous array. By doing
55 /// this, we reduce the number of allocations by a factor of B, and improve cache efficiency in
56 /// searches. However, this does mean that searches will have to do *more* comparisons on average.
57 /// The precise number of comparisons depends on the node search strategy used. For optimal cache
58 /// efficiency, one could search the nodes linearly. For optimal comparisons, one could search
59 /// the node using binary search. As a compromise, one could also perform a linear search
60 /// that initially only checks every i<sup>th</sup> element for some choice of i.
62 /// Currently, our implementation simply performs naive linear search. This provides excellent
63 /// performance on *small* nodes of elements which are cheap to compare. However in the future we
64 /// would like to further explore choosing the optimal search strategy based on the choice of B,
65 /// and possibly other factors. Using linear search, searching for a random element is expected
66 /// to take O(B log<sub>B</sub>n) comparisons, which is generally worse than a BST. In practice,
67 /// however, performance is excellent. `BTreeMap` is able to readily outperform `TreeMap` under
68 /// many workloads, and is competitive where it doesn't. BTreeMap also generally *scales* better
69 /// than TreeMap, making it more appropriate for large datasets.
71 /// However, `TreeMap` may still be more appropriate to use in many contexts. If elements are very
72 /// large or expensive to compare, `TreeMap` may be more appropriate. It won't allocate any
73 /// more space than is needed, and will perform the minimal number of comparisons necessary.
74 /// `TreeMap` also provides much better performance stability guarantees. Generally, very few
75 /// changes need to be made to update a BST, and two updates are expected to take about the same
76 /// amount of time on roughly equal sized BSTs. However a B-Tree's performance is much more
77 /// amortized. If a node is overfull, it must be split into two nodes. If a node is underfull, it
78 /// may be merged with another. Both of these operations are relatively expensive to perform, and
79 /// it's possible to force one to occur at every single level of the tree in a single insertion or
80 /// deletion. In fact, a malicious or otherwise unlucky sequence of insertions and deletions can
81 /// force this degenerate behaviour to occur on every operation. While the total amount of work
82 /// done on each operation isn't *catastrophic*, and *is* still bounded by O(B log<sub>B</sub>n),
83 /// it is certainly much slower when it does.
86 pub struct BTreeMap<K, V> {
93 /// An abstract base over-which all other BTree iterators are built.
101 /// An iterator over a BTreeMap's entries.
103 pub struct Iter<'a, K: 'a, V: 'a> {
104 inner: AbsIter<Traversal<'a, K, V>>
107 /// A mutable iterator over a BTreeMap's entries.
109 pub struct IterMut<'a, K: 'a, V: 'a> {
110 inner: AbsIter<MutTraversal<'a, K, V>>
113 /// An owning iterator over a BTreeMap's entries.
115 pub struct IntoIter<K, V> {
116 inner: AbsIter<MoveTraversal<K, V>>
119 /// An iterator over a BTreeMap's keys.
121 pub struct Keys<'a, K: 'a, V: 'a> {
122 inner: Map<(&'a K, &'a V), &'a K, Iter<'a, K, V>, fn((&'a K, &'a V)) -> &'a K>
125 /// An iterator over a BTreeMap's values.
127 pub struct Values<'a, K: 'a, V: 'a> {
128 inner: Map<(&'a K, &'a V), &'a V, Iter<'a, K, V>, fn((&'a K, &'a V)) -> &'a V>
131 /// A view into a single entry in a map, which may either be vacant or occupied.
132 pub enum Entry<'a, K:'a, V:'a> {
134 Vacant(VacantEntry<'a, K, V>),
135 /// An occupied Entry
136 Occupied(OccupiedEntry<'a, K, V>),
140 pub struct VacantEntry<'a, K:'a, V:'a> {
142 stack: stack::SearchStack<'a, K, V, node::handle::Edge, node::handle::Leaf>,
145 /// An occupied Entry.
146 pub struct OccupiedEntry<'a, K:'a, V:'a> {
147 stack: stack::SearchStack<'a, K, V, node::handle::KV, node::handle::LeafOrInternal>,
150 impl<K: Ord, V> BTreeMap<K, V> {
151 /// Makes a new empty BTreeMap with a reasonable choice for B.
153 pub fn new() -> BTreeMap<K, V> {
154 //FIXME(Gankro): Tune this as a function of size_of<K/V>?
158 /// Makes a new empty BTreeMap with the given B.
160 /// B cannot be less than 2.
161 pub fn with_b(b: uint) -> BTreeMap<K, V> {
162 assert!(b > 1, "B must be greater than 1");
166 root: Node::make_leaf_root(b),
171 /// Clears the map, removing all values.
176 /// use std::collections::BTreeMap;
178 /// let mut a = BTreeMap::new();
179 /// a.insert(1u, "a");
181 /// assert!(a.is_empty());
184 pub fn clear(&mut self) {
186 // avoid recursive destructors by manually traversing the tree
187 for _ in mem::replace(self, BTreeMap::with_b(b)).into_iter() {};
190 // Searching in a B-Tree is pretty straightforward.
192 // Start at the root. Try to find the key in the current node. If we find it, return it.
193 // If it's not in there, follow the edge *before* the smallest key larger than
194 // the search key. If no such key exists (they're *all* smaller), then just take the last
195 // edge in the node. If we're in a leaf and we don't find our key, then it's not
198 /// Returns a reference to the value corresponding to the key.
200 /// The key may be any borrowed form of the map's key type, but the ordering
201 /// on the borrowed form *must* match the ordering on the key type.
206 /// use std::collections::BTreeMap;
208 /// let mut map = BTreeMap::new();
209 /// map.insert(1u, "a");
210 /// assert_eq!(map.get(&1), Some(&"a"));
211 /// assert_eq!(map.get(&2), None);
214 pub fn get<Sized? Q>(&self, key: &Q) -> Option<&V> where Q: BorrowFrom<K> + Ord {
215 let mut cur_node = &self.root;
217 match Node::search(cur_node, key) {
218 Found(handle) => return Some(handle.into_kv().1),
219 GoDown(handle) => match handle.force() {
220 Leaf(_) => return None,
221 Internal(internal_handle) => {
222 cur_node = internal_handle.into_edge();
230 /// Returns true if the map contains a value for the specified key.
232 /// The key may be any borrowed form of the map's key type, but the ordering
233 /// on the borrowed form *must* match the ordering on the key type.
238 /// use std::collections::BTreeMap;
240 /// let mut map = BTreeMap::new();
241 /// map.insert(1u, "a");
242 /// assert_eq!(map.contains_key(&1), true);
243 /// assert_eq!(map.contains_key(&2), false);
246 pub fn contains_key<Sized? Q>(&self, key: &Q) -> bool where Q: BorrowFrom<K> + Ord {
247 self.get(key).is_some()
250 /// Returns a mutable reference to the value corresponding to the key.
252 /// The key may be any borrowed form of the map's key type, but the ordering
253 /// on the borrowed form *must* match the ordering on the key type.
258 /// use std::collections::BTreeMap;
260 /// let mut map = BTreeMap::new();
261 /// map.insert(1u, "a");
262 /// match map.get_mut(&1) {
263 /// Some(x) => *x = "b",
266 /// assert_eq!(map[1], "b");
268 // See `get` for implementation notes, this is basically a copy-paste with mut's added
270 pub fn get_mut<Sized? Q>(&mut self, key: &Q) -> Option<&mut V> where Q: BorrowFrom<K> + Ord {
271 // temp_node is a Borrowck hack for having a mutable value outlive a loop iteration
272 let mut temp_node = &mut self.root;
274 let cur_node = temp_node;
275 match Node::search(cur_node, key) {
276 Found(handle) => return Some(handle.into_kv_mut().1),
277 GoDown(handle) => match handle.force() {
278 Leaf(_) => return None,
279 Internal(internal_handle) => {
280 temp_node = internal_handle.into_edge_mut();
288 // Insertion in a B-Tree is a bit complicated.
290 // First we do the same kind of search described in `find`. But we need to maintain a stack of
291 // all the nodes/edges in our search path. If we find a match for the key we're trying to
292 // insert, just swap the vals and return the old ones. However, when we bottom out in a leaf,
293 // we attempt to insert our key-value pair at the same location we would want to follow another
296 // If the node has room, then this is done in the obvious way by shifting elements. However,
297 // if the node itself is full, we split node into two, and give its median key-value
298 // pair to its parent to insert the new node with. Of course, the parent may also be
299 // full, and insertion can propagate until we reach the root. If we reach the root, and
300 // it is *also* full, then we split the root and place the two nodes under a newly made root.
302 // Note that we subtly deviate from Open Data Structures in our implementation of split.
303 // ODS describes inserting into the node *regardless* of its capacity, and then
304 // splitting *afterwards* if it happens to be overfull. However, this is inefficient.
305 // Instead, we split beforehand, and then insert the key-value pair into the appropriate
306 // result node. This has two consequences:
308 // 1) While ODS produces a left node of size B-1, and a right node of size B,
309 // we may potentially reverse this. However, this shouldn't effect the analysis.
311 // 2) While ODS may potentially return the pair we *just* inserted after
312 // the split, we will never do this. Again, this shouldn't effect the analysis.
314 /// Inserts a key-value pair from the map. If the key already had a value
315 /// present in the map, that value is returned. Otherwise, `None` is returned.
320 /// use std::collections::BTreeMap;
322 /// let mut map = BTreeMap::new();
323 /// assert_eq!(map.insert(37u, "a"), None);
324 /// assert_eq!(map.is_empty(), false);
326 /// map.insert(37, "b");
327 /// assert_eq!(map.insert(37, "c"), Some("b"));
328 /// assert_eq!(map[37], "c");
331 pub fn insert(&mut self, mut key: K, mut value: V) -> Option<V> {
332 // This is a stack of rawptrs to nodes paired with indices, respectively
333 // representing the nodes and edges of our search path. We have to store rawptrs
334 // because as far as Rust is concerned, we can mutate aliased data with such a
335 // stack. It is of course correct, but what it doesn't know is that we will only
336 // be popping and using these ptrs one at a time in child-to-parent order. The alternative
337 // to doing this is to take the Nodes from their parents. This actually makes
338 // borrowck *really* happy and everything is pretty smooth. However, this creates
339 // *tons* of pointless writes, and requires us to always walk all the way back to
340 // the root after an insertion, even if we only needed to change a leaf. Therefore,
341 // we accept this potential unsafety and complexity in the name of performance.
343 // Regardless, the actual dangerous logic is completely abstracted away from BTreeMap
344 // by the stack module. All it can do is immutably read nodes, and ask the search stack
345 // to proceed down some edge by index. This makes the search logic we'll be reusing in a
346 // few different methods much neater, and of course drastically improves safety.
347 let mut stack = stack::PartialSearchStack::new(self);
350 let result = stack.with(move |pusher, node| {
351 // Same basic logic as found in `find`, but with PartialSearchStack mediating the
352 // actual nodes for us
353 return match Node::search(node, &key) {
354 Found(mut handle) => {
355 // Perfect match, swap the values and return the old one
356 mem::swap(handle.val_mut(), &mut value);
357 Finished(Some(value))
360 // We need to keep searching, try to get the search stack
361 // to go down further
362 match handle.force() {
363 Leaf(leaf_handle) => {
364 // We've reached a leaf, perform the insertion here
365 pusher.seal(leaf_handle).insert(key, value);
368 Internal(internal_handle) => {
369 // We've found the subtree to insert this key/value pair in,
371 Continue((pusher.push(internal_handle), key, value))
378 Finished(ret) => { return ret; },
379 Continue((new_stack, renewed_key, renewed_val)) => {
388 // Deletion is the most complicated operation for a B-Tree.
390 // First we do the same kind of search described in
391 // `find`. But we need to maintain a stack of all the nodes/edges in our search path.
392 // If we don't find the key, then we just return `None` and do nothing. If we do find the
393 // key, we perform two operations: remove the item, and then possibly handle underflow.
395 // # removing the item
396 // If the node is a leaf, we just remove the item, and shift
397 // any items after it back to fill the hole.
399 // If the node is an internal node, we *swap* the item with the smallest item in
400 // in its right subtree (which must reside in a leaf), and then revert to the leaf
403 // # handling underflow
404 // After removing an item, there may be too few items in the node. We want nodes
405 // to be mostly full for efficiency, although we make an exception for the root, which
406 // may have as few as one item. If this is the case, we may first try to steal
407 // an item from our left or right neighbour.
409 // To steal from the left (right) neighbour,
410 // we take the largest (smallest) item and child from it. We then swap the taken item
411 // with the item in their mutual parent that separates them, and then insert the
412 // parent's item and the taken child into the first (last) index of the underflowed node.
414 // However, stealing has the possibility of underflowing our neighbour. If this is the
415 // case, we instead *merge* with our neighbour. This of course reduces the number of
416 // children in the parent. Therefore, we also steal the item that separates the now
417 // merged nodes, and insert it into the merged node.
419 // Merging may cause the parent to underflow. If this is the case, then we must repeat
420 // the underflow handling process on the parent. If merging merges the last two children
421 // of the root, then we replace the root with the merged node.
423 /// Removes a key from the map, returning the value at the key if the key
424 /// was previously in the map.
426 /// The key may be any borrowed form of the map's key type, but the ordering
427 /// on the borrowed form *must* match the ordering on the key type.
432 /// use std::collections::BTreeMap;
434 /// let mut map = BTreeMap::new();
435 /// map.insert(1u, "a");
436 /// assert_eq!(map.remove(&1), Some("a"));
437 /// assert_eq!(map.remove(&1), None);
440 pub fn remove<Sized? Q>(&mut self, key: &Q) -> Option<V> where Q: BorrowFrom<K> + Ord {
441 // See `swap` for a more thorough description of the stuff going on in here
442 let mut stack = stack::PartialSearchStack::new(self);
444 let result = stack.with(move |pusher, node| {
445 return match Node::search(node, key) {
447 // Perfect match. Terminate the stack here, and remove the entry
448 Finished(Some(pusher.seal(handle).remove()))
451 // We need to keep searching, try to go down the next edge
452 match handle.force() {
453 // We're at a leaf; the key isn't in here
454 Leaf(_) => Finished(None),
455 Internal(internal_handle) => Continue(pusher.push(internal_handle))
461 Finished(ret) => return ret,
462 Continue(new_stack) => stack = new_stack
468 /// A helper enum useful for deciding whether to continue a loop since we can't
469 /// return from a closure
470 enum Continuation<A, B> {
475 /// The stack module provides a safe interface for constructing and manipulating a stack of ptrs
476 /// to nodes. By using this module much better safety guarantees can be made, and more search
477 /// boilerplate gets cut out.
479 use core::prelude::*;
480 use core::kinds::marker;
482 use core::ops::{Deref, DerefMut};
484 use super::super::node::{self, Node, Fit, Split, Internal, Leaf};
485 use super::super::node::handle;
488 /// A generic mutable reference, identical to `&mut` except for the fact that its lifetime
489 /// parameter is invariant. This means that wherever an `IdRef` is expected, only an `IdRef`
490 /// with the exact requested lifetime can be used. This is in contrast to normal references,
491 /// where `&'static` can be used in any function expecting any lifetime reference.
492 pub struct IdRef<'id, T: 'id> {
494 marker: marker::InvariantLifetime<'id>
497 impl<'id, T> Deref for IdRef<'id, T> {
500 fn deref(&self) -> &T {
505 impl<'id, T> DerefMut for IdRef<'id, T> {
506 fn deref_mut(&mut self) -> &mut T {
511 type StackItem<K, V> = node::Handle<*mut Node<K, V>, handle::Edge, handle::Internal>;
512 type Stack<K, V> = Vec<StackItem<K, V>>;
514 /// A `PartialSearchStack` handles the construction of a search stack.
515 pub struct PartialSearchStack<'a, K:'a, V:'a> {
516 map: &'a mut BTreeMap<K, V>,
518 next: *mut Node<K, V>,
521 /// A `SearchStack` represents a full path to an element or an edge of interest. It provides
522 /// methods depending on the type of what the path points to for removing an element, inserting
523 /// a new element, and manipulating to element at the top of the stack.
524 pub struct SearchStack<'a, K:'a, V:'a, Type, NodeType> {
525 map: &'a mut BTreeMap<K, V>,
527 top: node::Handle<*mut Node<K, V>, Type, NodeType>,
530 /// A `PartialSearchStack` that doesn't hold a a reference to the next node, and is just
531 /// just waiting for a `Handle` to that next node to be pushed. See `PartialSearchStack::with`
532 /// for more details.
533 pub struct Pusher<'id, 'a, K:'a, V:'a> {
534 map: &'a mut BTreeMap<K, V>,
536 marker: marker::InvariantLifetime<'id>
539 impl<'a, K, V> PartialSearchStack<'a, K, V> {
540 /// Creates a new PartialSearchStack from a BTreeMap by initializing the stack with the
541 /// root of the tree.
542 pub fn new(map: &'a mut BTreeMap<K, V>) -> PartialSearchStack<'a, K, V> {
543 let depth = map.depth;
546 next: &mut map.root as *mut _,
548 stack: Vec::with_capacity(depth),
552 /// Breaks up the stack into a `Pusher` and the next `Node`, allowing the given closure
553 /// to interact with, search, and finally push the `Node` onto the stack. The passed in
554 /// closure must be polymorphic on the `'id` lifetime parameter, as this statically
555 /// ensures that only `Handle`s from the correct `Node` can be pushed.
557 /// The reason this works is that the `Pusher` has an `'id` parameter, and will only accept
558 /// handles with the same `'id`. The closure could only get references with that lifetime
559 /// through its arguments or through some other `IdRef` that it has lying around. However,
560 /// no other `IdRef` could possibly work - because the `'id` is held in an invariant
561 /// parameter, it would need to have precisely the correct lifetime, which would mean that
562 /// at least one of the calls to `with` wouldn't be properly polymorphic, wanting a
563 /// specific lifetime instead of the one that `with` chooses to give it.
565 /// See also Haskell's `ST` monad, which uses a similar trick.
566 pub fn with<T, F: for<'id> FnOnce(Pusher<'id, 'a, K, V>,
567 IdRef<'id, Node<K, V>>) -> T>(self, closure: F) -> T {
568 let pusher = Pusher {
571 marker: marker::InvariantLifetime
574 inner: unsafe { &mut *self.next },
575 marker: marker::InvariantLifetime
578 closure(pusher, node)
582 impl<'id, 'a, K, V> Pusher<'id, 'a, K, V> {
583 /// Pushes the requested child of the stack's current top on top of the stack. If the child
584 /// exists, then a new PartialSearchStack is yielded. Otherwise, a VacantSearchStack is
586 pub fn push(mut self, mut edge: node::Handle<IdRef<'id, Node<K, V>>,
589 -> PartialSearchStack<'a, K, V> {
590 self.stack.push(edge.as_raw());
594 next: edge.edge_mut() as *mut _,
598 /// Converts the PartialSearchStack into a SearchStack.
599 pub fn seal<Type, NodeType>
600 (self, mut handle: node::Handle<IdRef<'id, Node<K, V>>, Type, NodeType>)
601 -> SearchStack<'a, K, V, Type, NodeType> {
605 top: handle.as_raw(),
610 impl<'a, K, V, NodeType> SearchStack<'a, K, V, handle::KV, NodeType> {
611 /// Gets a reference to the value the stack points to.
612 pub fn peek(&self) -> &V {
613 unsafe { self.top.from_raw().into_kv().1 }
616 /// Gets a mutable reference to the value the stack points to.
617 pub fn peek_mut(&mut self) -> &mut V {
618 unsafe { self.top.from_raw_mut().into_kv_mut().1 }
621 /// Converts the stack into a mutable reference to the value it points to, with a lifetime
622 /// tied to the original tree.
623 pub fn into_top(mut self) -> &'a mut V {
625 mem::copy_mut_lifetime(
627 self.top.from_raw_mut().val_mut()
633 impl<'a, K, V> SearchStack<'a, K, V, handle::KV, handle::Leaf> {
634 /// Removes the key and value in the top element of the stack, then handles underflows as
635 /// described in BTree's pop function.
636 fn remove_leaf(mut self) -> V {
637 self.map.length -= 1;
639 // Remove the key-value pair from the leaf that this search stack points to.
640 // Then, note if the leaf is underfull, and promptly forget the leaf and its ptr
641 // to avoid ownership issues.
642 let (value, mut underflow) = unsafe {
643 let (_, value) = self.top.from_raw_mut().remove_as_leaf();
644 let underflow = self.top.from_raw().node().is_underfull();
649 match self.stack.pop() {
651 // We've reached the root, so no matter what, we're done. We manually
652 // access the root via the tree itself to avoid creating any dangling
654 if self.map.root.len() == 0 && !self.map.root.is_leaf() {
655 // We've emptied out the root, so make its only child the new root.
656 // If it's a leaf, we just let it become empty.
658 self.map.root.hoist_lone_child();
662 Some(mut handle) => {
664 // Underflow! Handle it!
666 handle.from_raw_mut().handle_underflow();
667 underflow = handle.from_raw().node().is_underfull();
679 impl<'a, K, V> SearchStack<'a, K, V, handle::KV, handle::LeafOrInternal> {
680 /// Removes the key and value in the top element of the stack, then handles underflows as
681 /// described in BTree's pop function.
682 pub fn remove(self) -> V {
683 // Ensure that the search stack goes to a leaf. This is necessary to perform deletion
684 // in a BTree. Note that this may put the tree in an inconsistent state (further
685 // described in into_leaf's comments), but this is immediately fixed by the
686 // removing the value we want to remove
687 self.into_leaf().remove_leaf()
690 /// Subroutine for removal. Takes a search stack for a key that might terminate at an
691 /// internal node, and mutates the tree and search stack to *make* it a search stack
692 /// for that same key that *does* terminates at a leaf. If the mutation occurs, then this
693 /// leaves the tree in an inconsistent state that must be repaired by the caller by
694 /// removing the entry in question. Specifically the key-value pair and its successor will
696 fn into_leaf(mut self) -> SearchStack<'a, K, V, handle::KV, handle::Leaf> {
698 let mut top_raw = self.top;
699 let mut top = top_raw.from_raw_mut();
701 let key_ptr = top.key_mut() as *mut _;
702 let val_ptr = top.val_mut() as *mut _;
704 // Try to go into the right subtree of the found key to find its successor
706 Leaf(mut leaf_handle) => {
707 // We're a proper leaf stack, nothing to do
711 top: leaf_handle.as_raw()
714 Internal(mut internal_handle) => {
715 let mut right_handle = internal_handle.right_edge();
717 //We're not a proper leaf stack, let's get to work.
718 self.stack.push(right_handle.as_raw());
720 let mut temp_node = right_handle.edge_mut();
722 // Walk into the smallest subtree of this node
723 let node = temp_node;
725 match node.kv_handle(0).force() {
726 Leaf(mut handle) => {
727 // This node is a leaf, do the swap and return
728 mem::swap(handle.key_mut(), &mut *key_ptr);
729 mem::swap(handle.val_mut(), &mut *val_ptr);
736 Internal(kv_handle) => {
737 // This node is internal, go deeper
738 let mut handle = kv_handle.into_left_edge();
739 self.stack.push(handle.as_raw());
740 temp_node = handle.into_edge_mut();
750 impl<'a, K, V> SearchStack<'a, K, V, handle::Edge, handle::Leaf> {
751 /// Inserts the key and value into the top element in the stack, and if that node has to
752 /// split recursively inserts the split contents into the next element stack until
755 /// Assumes that the stack represents a search path from the root to a leaf.
757 /// An &mut V is returned to the inserted value, for callers that want a reference to this.
758 pub fn insert(mut self, key: K, val: V) -> &'a mut V {
760 self.map.length += 1;
762 // Insert the key and value into the leaf at the top of the stack
763 let (mut insertion, inserted_ptr) = self.top.from_raw_mut()
764 .insert_as_leaf(key, val);
769 // The last insertion went off without a hitch, no splits! We can stop
771 return &mut *inserted_ptr;
773 Split(key, val, right) => match self.stack.pop() {
774 // The last insertion triggered a split, so get the next element on the
775 // stack to recursively insert the split node into.
777 // The stack was empty; we've split the root, and need to make a
778 // a new one. This is done in-place because we can't move the
779 // root out of a reference to the tree.
780 Node::make_internal_root(&mut self.map.root, self.map.b,
784 return &mut *inserted_ptr;
786 Some(mut handle) => {
787 // The stack wasn't empty, do the insertion and recurse
788 insertion = handle.from_raw_mut()
789 .insert_as_internal(key, val, right);
801 impl<K: Ord, V> FromIterator<(K, V)> for BTreeMap<K, V> {
802 fn from_iter<T: Iterator<Item=(K, V)>>(iter: T) -> BTreeMap<K, V> {
803 let mut map = BTreeMap::new();
810 impl<K: Ord, V> Extend<(K, V)> for BTreeMap<K, V> {
812 fn extend<T: Iterator<Item=(K, V)>>(&mut self, mut iter: T) {
820 impl<S: Writer, K: Hash<S>, V: Hash<S>> Hash<S> for BTreeMap<K, V> {
821 fn hash(&self, state: &mut S) {
822 for elt in self.iter() {
829 impl<K: Ord, V> Default for BTreeMap<K, V> {
831 fn default() -> BTreeMap<K, V> {
837 impl<K: PartialEq, V: PartialEq> PartialEq for BTreeMap<K, V> {
838 fn eq(&self, other: &BTreeMap<K, V>) -> bool {
839 self.len() == other.len() &&
840 self.iter().zip(other.iter()).all(|(a, b)| a == b)
845 impl<K: Eq, V: Eq> Eq for BTreeMap<K, V> {}
848 impl<K: PartialOrd, V: PartialOrd> PartialOrd for BTreeMap<K, V> {
850 fn partial_cmp(&self, other: &BTreeMap<K, V>) -> Option<Ordering> {
851 iter::order::partial_cmp(self.iter(), other.iter())
856 impl<K: Ord, V: Ord> Ord for BTreeMap<K, V> {
858 fn cmp(&self, other: &BTreeMap<K, V>) -> Ordering {
859 iter::order::cmp(self.iter(), other.iter())
864 impl<K: Show, V: Show> Show for BTreeMap<K, V> {
865 fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
866 try!(write!(f, "{{"));
868 for (i, (k, v)) in self.iter().enumerate() {
869 if i != 0 { try!(write!(f, ", ")); }
870 try!(write!(f, "{}: {}", *k, *v));
877 // NOTE(stage0): remove impl after a snapshot
880 impl<K: Ord, Sized? Q, V> Index<Q, V> for BTreeMap<K, V>
881 where Q: BorrowFrom<K> + Ord
883 fn index(&self, key: &Q) -> &V {
884 self.get(key).expect("no entry found for key")
888 #[cfg(not(stage0))] // NOTE(stage0): remove cfg after a snapshot
890 impl<K: Ord, Sized? Q, V> Index<Q> for BTreeMap<K, V>
891 where Q: BorrowFrom<K> + Ord
895 fn index(&self, key: &Q) -> &V {
896 self.get(key).expect("no entry found for key")
900 // NOTE(stage0): remove impl after a snapshot
903 impl<K: Ord, Sized? Q, V> IndexMut<Q, V> for BTreeMap<K, V>
904 where Q: BorrowFrom<K> + Ord
906 fn index_mut(&mut self, key: &Q) -> &mut V {
907 self.get_mut(key).expect("no entry found for key")
911 #[cfg(not(stage0))] // NOTE(stage0): remove cfg after a snapshot
913 impl<K: Ord, Sized? Q, V> IndexMut<Q> for BTreeMap<K, V>
914 where Q: BorrowFrom<K> + Ord
918 fn index_mut(&mut self, key: &Q) -> &mut V {
919 self.get_mut(key).expect("no entry found for key")
923 /// Genericises over how to get the correct type of iterator from the correct type
924 /// of Node ownership.
926 fn traverse(node: N) -> Self;
929 impl<'a, K, V> Traverse<&'a Node<K, V>> for Traversal<'a, K, V> {
930 fn traverse(node: &'a Node<K, V>) -> Traversal<'a, K, V> {
935 impl<'a, K, V> Traverse<&'a mut Node<K, V>> for MutTraversal<'a, K, V> {
936 fn traverse(node: &'a mut Node<K, V>) -> MutTraversal<'a, K, V> {
941 impl<K, V> Traverse<Node<K, V>> for MoveTraversal<K, V> {
942 fn traverse(node: Node<K, V>) -> MoveTraversal<K, V> {
947 /// Represents an operation to perform inside the following iterator methods.
948 /// This is necessary to use in `next` because we want to modify self.left inside
949 /// a match that borrows it. Similarly, in `next_back` for self.right. Instead, we use this
950 /// enum to note what we want to do, and do it after the match.
956 impl<K, V, E, T> Iterator for AbsIter<T> where
957 T: DoubleEndedIterator + Iterator<Item=TraversalItem<K, V, E>> + Traverse<E>,
961 // This function is pretty long, but only because there's a lot of cases to consider.
962 // Our iterator represents two search paths, left and right, to the smallest and largest
963 // elements we have yet to yield. lca represents the least common ancestor of these two paths,
964 // above-which we never walk, since everything outside it has already been consumed (or was
965 // never in the range to iterate).
967 // Note that the design of these iterators permits an *arbitrary* initial pair of min and max,
968 // making these arbitrary sub-range iterators. However the logic to construct these paths
969 // efficiently is fairly involved, so this is a FIXME. The sub-range iterators also wouldn't be
970 // able to accurately predict size, so those iterators can't implement ExactSizeIterator.
971 fn next(&mut self) -> Option<(K, V)> {
973 // We want the smallest element, so try to get the top of the left stack
974 let op = match self.left.back_mut() {
975 // The left stack is empty, so try to get the next element of the two paths
976 // LCAs (the left search path is currently a subpath of the right one)
977 None => match self.lca.next() {
978 // The lca has been exhausted, walk further down the right path
979 None => match self.right.pop_front() {
980 // The right path is exhausted, so we're done
982 // The right path had something, make that the new LCA
983 // and restart the whole process
989 // The lca yielded an edge, make that the new head of the left path
990 Some(Edge(next)) => Push(Traverse::traverse(next)),
991 // The lca yielded an entry, so yield that
992 Some(Elem(k, v)) => {
997 // The left stack wasn't empty, so continue along the node in its head
998 Some(iter) => match iter.next() {
999 // The head of the left path is empty, so Pop it off and restart the process
1001 // The head of the left path yielded an edge, so make that the new head
1003 Some(Edge(next)) => Push(Traverse::traverse(next)),
1004 // The head of the left path yielded entry, so yield that
1005 Some(Elem(k, v)) => {
1012 // Handle any operation on the left stack as necessary
1014 Push(item) => { self.left.push_back(item); },
1015 Pop => { self.left.pop_back(); },
1020 fn size_hint(&self) -> (uint, Option<uint>) {
1021 (self.size, Some(self.size))
1025 impl<K, V, E, T> DoubleEndedIterator for AbsIter<T> where
1026 T: DoubleEndedIterator + Iterator<Item=TraversalItem<K, V, E>> + Traverse<E>,
1028 // next_back is totally symmetric to next
1029 fn next_back(&mut self) -> Option<(K, V)> {
1031 let op = match self.right.back_mut() {
1032 None => match self.lca.next_back() {
1033 None => match self.left.pop_front() {
1034 None => return None,
1040 Some(Edge(next)) => Push(Traverse::traverse(next)),
1041 Some(Elem(k, v)) => {
1046 Some(iter) => match iter.next_back() {
1048 Some(Edge(next)) => Push(Traverse::traverse(next)),
1049 Some(Elem(k, v)) => {
1057 Push(item) => { self.right.push_back(item); },
1058 Pop => { self.right.pop_back(); }
1065 impl<'a, K, V> Iterator for Iter<'a, K, V> {
1066 type Item = (&'a K, &'a V);
1068 fn next(&mut self) -> Option<(&'a K, &'a V)> { self.inner.next() }
1069 fn size_hint(&self) -> (uint, Option<uint>) { self.inner.size_hint() }
1072 impl<'a, K, V> DoubleEndedIterator for Iter<'a, K, V> {
1073 fn next_back(&mut self) -> Option<(&'a K, &'a V)> { self.inner.next_back() }
1076 impl<'a, K, V> ExactSizeIterator for Iter<'a, K, V> {}
1079 impl<'a, K, V> Iterator for IterMut<'a, K, V> {
1080 type Item = (&'a K, &'a mut V);
1082 fn next(&mut self) -> Option<(&'a K, &'a mut V)> { self.inner.next() }
1083 fn size_hint(&self) -> (uint, Option<uint>) { self.inner.size_hint() }
1086 impl<'a, K, V> DoubleEndedIterator for IterMut<'a, K, V> {
1087 fn next_back(&mut self) -> Option<(&'a K, &'a mut V)> { self.inner.next_back() }
1090 impl<'a, K, V> ExactSizeIterator for IterMut<'a, K, V> {}
1093 impl<K, V> Iterator for IntoIter<K, V> {
1096 fn next(&mut self) -> Option<(K, V)> { self.inner.next() }
1097 fn size_hint(&self) -> (uint, Option<uint>) { self.inner.size_hint() }
1100 impl<K, V> DoubleEndedIterator for IntoIter<K, V> {
1101 fn next_back(&mut self) -> Option<(K, V)> { self.inner.next_back() }
1104 impl<K, V> ExactSizeIterator for IntoIter<K, V> {}
1107 impl<'a, K, V> Iterator for Keys<'a, K, V> {
1110 fn next(&mut self) -> Option<(&'a K)> { self.inner.next() }
1111 fn size_hint(&self) -> (uint, Option<uint>) { self.inner.size_hint() }
1114 impl<'a, K, V> DoubleEndedIterator for Keys<'a, K, V> {
1115 fn next_back(&mut self) -> Option<(&'a K)> { self.inner.next_back() }
1118 impl<'a, K, V> ExactSizeIterator for Keys<'a, K, V> {}
1122 impl<'a, K, V> Iterator for Values<'a, K, V> {
1125 fn next(&mut self) -> Option<(&'a V)> { self.inner.next() }
1126 fn size_hint(&self) -> (uint, Option<uint>) { self.inner.size_hint() }
1129 impl<'a, K, V> DoubleEndedIterator for Values<'a, K, V> {
1130 fn next_back(&mut self) -> Option<(&'a V)> { self.inner.next_back() }
1133 impl<'a, K, V> ExactSizeIterator for Values<'a, K, V> {}
1136 impl<'a, K: Ord, V> VacantEntry<'a, K, V> {
1137 /// Sets the value of the entry with the VacantEntry's key,
1138 /// and returns a mutable reference to it.
1139 pub fn set(self, value: V) -> &'a mut V {
1140 self.stack.insert(self.key, value)
1144 impl<'a, K: Ord, V> OccupiedEntry<'a, K, V> {
1145 /// Gets a reference to the value in the entry.
1146 pub fn get(&self) -> &V {
1150 /// Gets a mutable reference to the value in the entry.
1151 pub fn get_mut(&mut self) -> &mut V {
1152 self.stack.peek_mut()
1155 /// Converts the entry into a mutable reference to its value.
1156 pub fn into_mut(self) -> &'a mut V {
1157 self.stack.into_top()
1160 /// Sets the value of the entry with the OccupiedEntry's key,
1161 /// and returns the entry's old value.
1162 pub fn set(&mut self, mut value: V) -> V {
1163 mem::swap(self.stack.peek_mut(), &mut value);
1167 /// Takes the value of the entry out of the map, and returns it.
1168 pub fn take(self) -> V {
1173 impl<K, V> BTreeMap<K, V> {
1174 /// Gets an iterator over the entries of the map.
1179 /// use std::collections::BTreeMap;
1181 /// let mut map = BTreeMap::new();
1182 /// map.insert(1u, "a");
1183 /// map.insert(2u, "b");
1184 /// map.insert(3u, "c");
1186 /// for (key, value) in map.iter() {
1187 /// println!("{}: {}", key, value);
1190 /// let (first_key, first_value) = map.iter().next().unwrap();
1191 /// assert_eq!((*first_key, *first_value), (1u, "a"));
1194 pub fn iter(&self) -> Iter<K, V> {
1195 let len = self.len();
1198 lca: Traverse::traverse(&self.root),
1199 left: RingBuf::new(),
1200 right: RingBuf::new(),
1206 /// Gets a mutable iterator over the entries of the map.
1211 /// use std::collections::BTreeMap;
1213 /// let mut map = BTreeMap::new();
1214 /// map.insert("a", 1u);
1215 /// map.insert("b", 2u);
1216 /// map.insert("c", 3u);
1218 /// // add 10 to the value if the key isn't "a"
1219 /// for (key, value) in map.iter_mut() {
1220 /// if key != &"a" {
1226 pub fn iter_mut(&mut self) -> IterMut<K, V> {
1227 let len = self.len();
1230 lca: Traverse::traverse(&mut self.root),
1231 left: RingBuf::new(),
1232 right: RingBuf::new(),
1238 /// Gets an owning iterator over the entries of the map.
1243 /// use std::collections::BTreeMap;
1245 /// let mut map = BTreeMap::new();
1246 /// map.insert(1u, "a");
1247 /// map.insert(2u, "b");
1248 /// map.insert(3u, "c");
1250 /// for (key, value) in map.into_iter() {
1251 /// println!("{}: {}", key, value);
1255 pub fn into_iter(self) -> IntoIter<K, V> {
1256 let len = self.len();
1259 lca: Traverse::traverse(self.root),
1260 left: RingBuf::new(),
1261 right: RingBuf::new(),
1267 /// Gets an iterator over the keys of the map.
1272 /// use std::collections::BTreeMap;
1274 /// let mut a = BTreeMap::new();
1275 /// a.insert(1u, "a");
1276 /// a.insert(2u, "b");
1278 /// let keys: Vec<uint> = a.keys().cloned().collect();
1279 /// assert_eq!(keys, vec![1u,2,]);
1282 pub fn keys<'a>(&'a self) -> Keys<'a, K, V> {
1283 fn first<A, B>((a, _): (A, B)) -> A { a }
1284 let first: fn((&'a K, &'a V)) -> &'a K = first; // coerce to fn pointer
1286 Keys { inner: self.iter().map(first) }
1289 /// Gets an iterator over the values of the map.
1294 /// use std::collections::BTreeMap;
1296 /// let mut a = BTreeMap::new();
1297 /// a.insert(1u, "a");
1298 /// a.insert(2u, "b");
1300 /// let values: Vec<&str> = a.values().cloned().collect();
1301 /// assert_eq!(values, vec!["a","b"]);
1304 pub fn values<'a>(&'a self) -> Values<'a, K, V> {
1305 fn second<A, B>((_, b): (A, B)) -> B { b }
1306 let second: fn((&'a K, &'a V)) -> &'a V = second; // coerce to fn pointer
1308 Values { inner: self.iter().map(second) }
1311 /// Return the number of elements in the map.
1316 /// use std::collections::BTreeMap;
1318 /// let mut a = BTreeMap::new();
1319 /// assert_eq!(a.len(), 0);
1320 /// a.insert(1u, "a");
1321 /// assert_eq!(a.len(), 1);
1324 pub fn len(&self) -> uint { self.length }
1326 /// Return true if the map contains no elements.
1331 /// use std::collections::BTreeMap;
1333 /// let mut a = BTreeMap::new();
1334 /// assert!(a.is_empty());
1335 /// a.insert(1u, "a");
1336 /// assert!(!a.is_empty());
1339 pub fn is_empty(&self) -> bool { self.len() == 0 }
1342 impl<K: Ord, V> BTreeMap<K, V> {
1343 /// Gets the given key's corresponding entry in the map for in-place manipulation.
1348 /// use std::collections::BTreeMap;
1349 /// use std::collections::btree_map::Entry;
1351 /// let mut count: BTreeMap<&str, uint> = BTreeMap::new();
1353 /// // count the number of occurrences of letters in the vec
1354 /// for x in vec!["a","b","a","c","a","b"].iter() {
1355 /// match count.entry(*x) {
1356 /// Entry::Vacant(view) => {
1359 /// Entry::Occupied(mut view) => {
1360 /// let v = view.get_mut();
1366 /// assert_eq!(count["a"], 3u);
1368 pub fn entry<'a>(&'a mut self, mut key: K) -> Entry<'a, K, V> {
1369 // same basic logic of `swap` and `pop`, blended together
1370 let mut stack = stack::PartialSearchStack::new(self);
1372 let result = stack.with(move |pusher, node| {
1373 return match Node::search(node, &key) {
1376 Finished(Occupied(OccupiedEntry {
1377 stack: pusher.seal(handle)
1381 match handle.force() {
1382 Leaf(leaf_handle) => {
1383 Finished(Vacant(VacantEntry {
1384 stack: pusher.seal(leaf_handle),
1388 Internal(internal_handle) => {
1390 pusher.push(internal_handle),
1399 Finished(finished) => return finished,
1400 Continue((new_stack, renewed_key)) => {
1417 use super::{BTreeMap, Occupied, Vacant};
1420 fn test_basic_large() {
1421 let mut map = BTreeMap::new();
1423 assert_eq!(map.len(), 0);
1425 for i in range(0, size) {
1426 assert_eq!(map.insert(i, 10*i), None);
1427 assert_eq!(map.len(), i + 1);
1430 for i in range(0, size) {
1431 assert_eq!(map.get(&i).unwrap(), &(i*10));
1434 for i in range(size, size*2) {
1435 assert_eq!(map.get(&i), None);
1438 for i in range(0, size) {
1439 assert_eq!(map.insert(i, 100*i), Some(10*i));
1440 assert_eq!(map.len(), size);
1443 for i in range(0, size) {
1444 assert_eq!(map.get(&i).unwrap(), &(i*100));
1447 for i in range(0, size/2) {
1448 assert_eq!(map.remove(&(i*2)), Some(i*200));
1449 assert_eq!(map.len(), size - i - 1);
1452 for i in range(0, size/2) {
1453 assert_eq!(map.get(&(2*i)), None);
1454 assert_eq!(map.get(&(2*i+1)).unwrap(), &(i*200 + 100));
1457 for i in range(0, size/2) {
1458 assert_eq!(map.remove(&(2*i)), None);
1459 assert_eq!(map.remove(&(2*i+1)), Some(i*200 + 100));
1460 assert_eq!(map.len(), size/2 - i - 1);
1465 fn test_basic_small() {
1466 let mut map = BTreeMap::new();
1467 assert_eq!(map.remove(&1), None);
1468 assert_eq!(map.get(&1), None);
1469 assert_eq!(map.insert(1u, 1u), None);
1470 assert_eq!(map.get(&1), Some(&1));
1471 assert_eq!(map.insert(1, 2), Some(1));
1472 assert_eq!(map.get(&1), Some(&2));
1473 assert_eq!(map.insert(2, 4), None);
1474 assert_eq!(map.get(&2), Some(&4));
1475 assert_eq!(map.remove(&1), Some(2));
1476 assert_eq!(map.remove(&2), Some(4));
1477 assert_eq!(map.remove(&1), None);
1485 let mut map: BTreeMap<uint, uint> = range(0, size).map(|i| (i, i)).collect();
1488 let mut iter = map.iter();
1489 for i in range(0, size) {
1490 assert_eq!(iter.size_hint(), (size - i, Some(size - i)));
1491 assert_eq!(iter.next().unwrap(), (&i, &i));
1493 assert_eq!(iter.size_hint(), (0, Some(0)));
1494 assert_eq!(iter.next(), None);
1498 let mut iter = map.iter_mut();
1499 for i in range(0, size) {
1500 assert_eq!(iter.size_hint(), (size - i, Some(size - i)));
1501 assert_eq!(iter.next().unwrap(), (&i, &mut (i + 0)));
1503 assert_eq!(iter.size_hint(), (0, Some(0)));
1504 assert_eq!(iter.next(), None);
1508 let mut iter = map.into_iter();
1509 for i in range(0, size) {
1510 assert_eq!(iter.size_hint(), (size - i, Some(size - i)));
1511 assert_eq!(iter.next().unwrap(), (i, i));
1513 assert_eq!(iter.size_hint(), (0, Some(0)));
1514 assert_eq!(iter.next(), None);
1520 fn test_iter_rev() {
1524 let mut map: BTreeMap<uint, uint> = range(0, size).map(|i| (i, i)).collect();
1527 let mut iter = map.iter().rev();
1528 for i in range(0, size) {
1529 assert_eq!(iter.size_hint(), (size - i, Some(size - i)));
1530 assert_eq!(iter.next().unwrap(), (&(size - i - 1), &(size - i - 1)));
1532 assert_eq!(iter.size_hint(), (0, Some(0)));
1533 assert_eq!(iter.next(), None);
1537 let mut iter = map.iter_mut().rev();
1538 for i in range(0, size) {
1539 assert_eq!(iter.size_hint(), (size - i, Some(size - i)));
1540 assert_eq!(iter.next().unwrap(), (&(size - i - 1), &mut(size - i - 1)));
1542 assert_eq!(iter.size_hint(), (0, Some(0)));
1543 assert_eq!(iter.next(), None);
1547 let mut iter = map.into_iter().rev();
1548 for i in range(0, size) {
1549 assert_eq!(iter.size_hint(), (size - i, Some(size - i)));
1550 assert_eq!(iter.next().unwrap(), (size - i - 1, size - i - 1));
1552 assert_eq!(iter.size_hint(), (0, Some(0)));
1553 assert_eq!(iter.next(), None);
1560 let xs = [(1i, 10i), (2, 20), (3, 30), (4, 40), (5, 50), (6, 60)];
1562 let mut map: BTreeMap<int, int> = xs.iter().map(|&x| x).collect();
1564 // Existing key (insert)
1565 match map.entry(1) {
1566 Vacant(_) => unreachable!(),
1567 Occupied(mut view) => {
1568 assert_eq!(view.get(), &10);
1569 assert_eq!(view.set(100), 10);
1572 assert_eq!(map.get(&1).unwrap(), &100);
1573 assert_eq!(map.len(), 6);
1576 // Existing key (update)
1577 match map.entry(2) {
1578 Vacant(_) => unreachable!(),
1579 Occupied(mut view) => {
1580 let v = view.get_mut();
1584 assert_eq!(map.get(&2).unwrap(), &200);
1585 assert_eq!(map.len(), 6);
1587 // Existing key (take)
1588 match map.entry(3) {
1589 Vacant(_) => unreachable!(),
1591 assert_eq!(view.take(), 30);
1594 assert_eq!(map.get(&3), None);
1595 assert_eq!(map.len(), 5);
1598 // Inexistent key (insert)
1599 match map.entry(10) {
1600 Occupied(_) => unreachable!(),
1602 assert_eq!(*view.set(1000), 1000);
1605 assert_eq!(map.get(&10).unwrap(), &1000);
1606 assert_eq!(map.len(), 6);
1618 use std::rand::{weak_rng, Rng};
1619 use test::{Bencher, black_box};
1621 use super::BTreeMap;
1622 use bench::{insert_rand_n, insert_seq_n, find_rand_n, find_seq_n};
1625 pub fn insert_rand_100(b: &mut Bencher) {
1626 let mut m : BTreeMap<uint,uint> = BTreeMap::new();
1627 insert_rand_n(100, &mut m, b,
1628 |m, i| { m.insert(i, 1); },
1629 |m, i| { m.remove(&i); });
1633 pub fn insert_rand_10_000(b: &mut Bencher) {
1634 let mut m : BTreeMap<uint,uint> = BTreeMap::new();
1635 insert_rand_n(10_000, &mut m, b,
1636 |m, i| { m.insert(i, 1); },
1637 |m, i| { m.remove(&i); });
1642 pub fn insert_seq_100(b: &mut Bencher) {
1643 let mut m : BTreeMap<uint,uint> = BTreeMap::new();
1644 insert_seq_n(100, &mut m, b,
1645 |m, i| { m.insert(i, 1); },
1646 |m, i| { m.remove(&i); });
1650 pub fn insert_seq_10_000(b: &mut Bencher) {
1651 let mut m : BTreeMap<uint,uint> = BTreeMap::new();
1652 insert_seq_n(10_000, &mut m, b,
1653 |m, i| { m.insert(i, 1); },
1654 |m, i| { m.remove(&i); });
1659 pub fn find_rand_100(b: &mut Bencher) {
1660 let mut m : BTreeMap<uint,uint> = BTreeMap::new();
1661 find_rand_n(100, &mut m, b,
1662 |m, i| { m.insert(i, 1); },
1663 |m, i| { m.get(&i); });
1667 pub fn find_rand_10_000(b: &mut Bencher) {
1668 let mut m : BTreeMap<uint,uint> = BTreeMap::new();
1669 find_rand_n(10_000, &mut m, b,
1670 |m, i| { m.insert(i, 1); },
1671 |m, i| { m.get(&i); });
1676 pub fn find_seq_100(b: &mut Bencher) {
1677 let mut m : BTreeMap<uint,uint> = BTreeMap::new();
1678 find_seq_n(100, &mut m, b,
1679 |m, i| { m.insert(i, 1); },
1680 |m, i| { m.get(&i); });
1684 pub fn find_seq_10_000(b: &mut Bencher) {
1685 let mut m : BTreeMap<uint,uint> = BTreeMap::new();
1686 find_seq_n(10_000, &mut m, b,
1687 |m, i| { m.insert(i, 1); },
1688 |m, i| { m.get(&i); });
1691 fn bench_iter(b: &mut Bencher, size: uint) {
1692 let mut map = BTreeMap::<uint, uint>::new();
1693 let mut rng = weak_rng();
1695 for _ in range(0, size) {
1696 map.insert(rng.gen(), rng.gen());
1700 for entry in map.iter() {
1707 pub fn iter_20(b: &mut Bencher) {
1712 pub fn iter_1000(b: &mut Bencher) {
1713 bench_iter(b, 1000);
1717 pub fn iter_100000(b: &mut Bencher) {
1718 bench_iter(b, 100000);