Boost.MultiIndex Hashed indices reference

Ordered indices

Boost.MultiIndex reference

Sequenced indices

Contents

• Header "boost/multi_index/hashed_index_fwd.hpp" synopsis
• Header "boost/multi_index/hashed_index.hpp" synopsis
  • Index specifiers hashed_unique and hashed_non_unique
  • Hashed indices
    • Complexity signature
    • Instantiation types
    • Nested types
    • Constructors, copy and assignment
    • Modifiers
    • Observers
    • Lookup
    • Hash policy
    • Serialization

Header "boost/multi_index/hashed_index_fwd.hpp" synopsis

namespace boost{

namespace multi_index{

// index specifiers hashed_unique and hashed_non_unique

template<consult hashed_unique reference for arguments>
struct hashed_unique;
template<consult hashed_non_unique reference for arguments>
struct hashed_non_unique;

// indices

namespace detail{

template<implementation defined> class index name is implementation defined;

} // namespace boost::multi_index::detail

} // namespace boost::multi_index 

} // namespace boost

hashed_index_fwd.hpp provides forward declarations for index specifiers hashed_unique and hashed_non_unique and their associated hashed index classes.

Header "boost/multi_index/hashed_index.hpp" synopsis

namespace boost{

namespace multi_index{

// index specifiers hashed_unique and hashed_non_unique

template<consult hashed_unique reference for arguments>
struct hashed_unique;
template<consult hashed_non_unique reference for arguments>
struct hashed_non_unique;

// indices

namespace detail{

template<implementation defined> class index class name implementation defined;

// index specialized algorithms:

template<implementation defined>
void swap(index class name& x,index class name& y);

} // namespace boost::multi_index::detail

} // namespace boost::multi_index 

} // namespace boost

Index specifiers hashed_unique and hashed_non_unique

These index specifiers allow for insertion of hashed indices without and with allowance of duplicate elements, respectively. The syntax of hashed_unique and hashed_non_unique coincide, thus we describe them in a grouped manner. hashed_unique and hashed_non_unique can be instantiated in two different forms, according to whether a tag list for the index is provided or not:

template<
  typename KeyFromValue,
  typename Hash=boost::hash<KeyFromValue::result_type>,
  typename Pred=std::equal_to<KeyFromValue::result_type>
>
struct (hashed_unique | hashed_non_unique);

template<
  typename TagList,
  typename KeyFromValue,
  typename Hash=boost::hash<KeyFromValue::result_type>,
  typename Pred=std::equal_to<KeyFromValue::result_type>
>
struct (hashed_unique | hashed_non_unique);

If provided, TagList must be an instantiation of the class template tag. The template arguments are used by the corresponding index implementation, refer to the hashed indices reference section for further explanations on their acceptable type values.

Hashed indices

A hashed index provides fast retrieval of elements of a multi_index_container through hashing tecnhiques. The interface and semantics of hashed indices are modeled according to the proposal for unordered associative containers given in the C++ Proposed Draft Tecnhical Report on Standard Library Extensions, also known as TR1. A hashed index is particularized according to a given Key Extractor that retrieves keys from elements of multi_index_container, a Hash function object which returns hash values for the keys and a binary predicate Pred acting as an equivalence relation on values of Key.

There are two variants of hashed indices: unique, which do not allow duplicate elements (with respect to its associated equality predicate) and non-unique, which accept those duplicates. The interface of these two variants is the same, so they are documented together, with minor differences explicitly stated when they exist.

Except where noted, hashed indices (both unique and non-unique) are models of Unordered Associative Container, in the spirit of std::tr1::unordered_sets. Validity of iterators and references to elements is preserved in all cases. Occasionally, the exception safety guarantees provided are actually stronger than required by the extension draft. We only provide descriptions of those types and operations that are either not present in the concepts modeled or do not exactly conform to the requirements for unordered associative containers.

namespace boost{

namespace multi_index{

namespace detail{

template<implementation defined: dependent on types Value, Allocator,
  TagList, KeyFromValue, Hash, Pred>
class name is implementation defined
{ 
public:
  // types:

  typedef typename KeyFromValue::result_type         key_type;
  typedef Value                                      value_type;
  typedef KeyFromValue                               key_from_value;
  typedef Hash                                       hasher;
  typedef Pred                                       key_equal;
  typedef tuple<
    size_type,key_from_value,hasher,key_equal>       ctor_args;
  typedef Allocator                                  allocator_type;
  typedef typename Allocator::pointer                pointer;
  typedef typename Allocator::const_pointer          const_pointer;
  typedef typename Allocator::reference              reference;
  typedef typename Allocator::const_reference        const_reference;
  typedef implementation defined                     size_type;      
  typedef implementation defined                     difference_type;
  typedef implementation defined                     iterator;
  typedef implementation defined                     const_iterator;
  typedef implementation defined                     local_iterator;
  typedef implementation defined                     const_local_iterator;

  // construct/destroy/copy:

  index class name& operator=(const index class name& x);

  allocator_type get_allocator()const;

  // size and capacity:

  bool      empty()const;
  size_type size()const;
  size_type max_size()const;

  // iterators:

  iterator       begin();
  const_iterator begin()const;
  iterator       end();
  const_iterator end()const;
 
  // modifiers:

  std::pair<iterator,bool> insert(const value_type& x);
  iterator insert(iterator position,const value_type& x);
  template<typename InputIterator>
  void insert(InputIterator first,InputIterator last);

  iterator  erase(iterator position);
  size_type erase(const key_type& x);
  iterator  erase(iterator first,iterator last);

  bool replace(iterator position,const value_type& x);
  template<typename Modifier> bool modify(iterator position,Modifier mod);
  template<typename Modifier> bool modify_key(iterator position,Modifier mod);
  
  void clear();
  void swap(index class name& x);

  // observers:

  key_from_value key_extractor()const;
  hasher         hash_function()const;
  key_equal      key_eq()const;

  // lookup:

  template<typename CompatibleKey>
  iterator find(const CompatibleKey& x)const;
  template<
    typename CompatibleKey,typename CompatibleHash, typename CompatiblePred
  >
  iterator find(
    const CompatibleKey& x,
    const CompatibleHash& hash,const CompatiblePred& eq)const; 

  template<typename CompatibleKey>
  size_type count(const CompatibleKey& x)const;
  template<
    typename CompatibleKey,typename CompatibleHash, typename CompatiblePred
  >
  size_type count(
    const CompatibleKey& x,
    const CompatibleHash& hash,const CompatiblePred& eq)const;

  template<typename CompatibleKey>
  std::pair<iterator,iterator> equal_range(const CompatibleKey& x)const;
  template<
    typename CompatibleKey,typename CompatibleHash, typename CompatiblePred
  >
  std::pair<iterator,iterator> equal_range(
    const CompatibleKey& x,
    const CompatibleHash& hash,const CompatiblePred& eq)const;

  // bucket interface:

  size_type bucket_count()const;
  size_type max_bucket_count()const;
  size_type bucket_size(size_type n)const;
  size_type bucket(const key_type& k)const;

  local_iterator       begin(size_type n);
  const_local_iterator begin(size_type n)const;
  local_iterator       end(size_type n);
  const_local_iterator end(size_type n)const;

  // hash policy:

  float load_factor()const;
  float max_load_factor()const;
  void  max_load_factor(float z);
  void  rehash(size_type n);
};

// index specialized algorithms:

template<implementation defined>
void swap(index class name& x,index class name& y);

} // namespace boost::multi_index::detail

} // namespace boost::multi_index 

} // namespace boost

Complexity signature

Here and in the descriptions of operations of hashed indices, we adopt the scheme outlined in the complexity signature section. The complexity signature of hashed indices is:

• copying: c(n)=n*log(n),
• insertion: average case i(n)=1 (constant), worst case i(n)=n,
• hinted insertion: average case h(n)=1 (constant), worst case h(n)=n,
• deletion: average case d(n)=1 (constant), worst case d(n)=n,
• replacement:
  • if the new element key is equivalent to the original, r(n)=1 (constant),
  • otherwise, average case r(n)=1 (constant), worst case r(n)=n,
• modifying: average case m(n)=1 (constant), worst case m(n)=n.

Instantiation types

Hashed indices are instantiated internally to multi_index_container and specified by means of indexed_by with index specifiers hashed_unique and hashed_non_unique. Instantiations are dependent on the following types:

• Value from multi_index_container,
• Allocator from multi_index_container,
• TagList from the index specifier (if provided),
• KeyFromValue from the index specifier,
• Hash from the index specifier,
• Pred from the index specifier.

Nested types

The first element of this tuple indicates the minimum number of buckets set up by the index on construction time. If the default value 0 is used, an implementation defined number is used instead.

These types are models of Forward Iterator.

Constructors, copy and assignment

As explained in the index concepts section, indices do not have public constructors or destructors. Assignment, on the other hand, is provided. Upon construction, max_load_factor() is 1.0.

Effects:

a=b;

where a and b are the multi_index_container objects to which *this and x belong, respectively.
Returns: *this.

Modifiers

Effects: Inserts x into the multi_index_container to which the index belongs if

• the index is non-unique OR no other element exists with equivalent key,
• AND insertion is allowed by all other indices of the multi_index_container.

Returns: The return value is a pair p. p.second is true if and only if insertion took place. On successful insertion, p.first points to the element inserted; otherwise, p.first points to an element that caused the insertion to be banned. Note that more than one element can be causing insertion not to be allowed.
Complexity: O(I(n)).
Exception safety: Strong.

Requires: position is a valid iterator of the index.
Effects: Inserts x into the multi_index_container to which the index belongs if

• the index is non-unique OR no other element exists with equivalent key,
• AND insertion is allowed by all other indices of the multi_index_container.

position is used as a hint to improve the efficiency of the operation.
Returns: On successful insertion, an iterator to the newly inserted element. Otherwise, an iterator to an element that caused the insertion to be banned. Note that more than one element can be causing insertion not to be allowed.
Complexity: O(H(n)).
Exception safety: Strong.

Requires: InputIterator is a model of Input Iterator over elements of type value_type or a type convertible to value_type. first and last are not iterators into any index of the multi_index_container to which this index belongs. last is reachable from first.
Effects:

iterator hint=end();
while(first!=last)hint=insert(hint,*first++);

Complexity: O(m*H(n+m)), where m is the number of elements in [first, last).
Exception safety: Basic.

Requires: position is a valid dereferenceable iterator of the index.
Effects: Deletes the element pointed to by position.
Returns: An iterator pointing to the element immediately following the one that was deleted, or end() if no such element exists.
Complexity: O(D(n)).
Exception safety: nothrow.

Effects: Deletes the elements with key equivalent to x.
Returns: Number of elements deleted.
Complexity: Average case, O(1 + m*D(n)), worst case O(n + m*D(n)), where m is the number of elements deleted.
Exception safety: Basic.

Requires: [first,last) is a valid range of the index.
Effects: Deletes the elements in [first,last).
Returns: last.
Complexity: O(m*D(n)), where m is the number of elements in [first,last).
Exception safety: nothrow.

Requires: position is a valid dereferenceable iterator of the index.
Effects: Assigns the value x to the element pointed to by position into the multi_index_container to which the index belongs if, for the value x

• the index is non-unique OR no other element exists (except possibly *position) with equivalent key,
• AND replacing is allowed by all other indices of the multi_index_container.

Postconditions: Validity of position is preserved in all cases.
Returns: true if the replacement took place, false otherwise.
Complexity: O(R(n)).
Exception safety: Strong. If an exception is thrown by some user-provided operation the multi_index_container to which the index belongs remains in its original state.

Requires: Modifier is a model of Unary Function accepting arguments of type value_type&. position is a valid dereferenceable iterator of the index.
Effects: Calls mod(e) where e is the element pointed to by position and rearranges *position into all the indices of the multi_index_container. Rearrangement is successful if

• the index is non-unique OR no other element exists with equivalent key,
• AND rearrangement is allowed by all other indices of the multi_index_container.

If the rearrangement fails, the element is erased.
Postconditions: Validity of position is preserved if the operation succeeds.
Returns: true if the operation succeeded, false otherwise.
Complexity: O(M(n)).
Exception safety: Basic. If an exception is thrown by some user-provided operation (except possibly mod), then the element pointed to by position is erased.

Requires: key_from_value is a read/write Key Extractor from value_type. Modifier is a model of Unary Function accepting arguments of type key_type&. position is a valid dereferenceable iterator of the index.
Effects: Calls mod(k) where k is the key obtained by the internal KeyFromValue object of the index from the element pointed to by position, and rearranges *position into all the indices of the multi_index_container. Rearrangement is successful if

• the index is non-unique OR no other element exists with equivalent key,
• AND rearrangement is allowed by all other indices of the multi_index_container.

If the rearrangement fails, the element is erased.
Postconditions:Validity of position is preserved if the operation succeeds.
Returns: true if the operation succeeded, false otherwise.
Complexity: O(M(n)).
Exception safety: Basic. If an exception is thrown by some user-provided operation (except possibly mod), then the element pointed to by position is erased.

Observers

Apart from standard hash_function and key_eq, hashed indices have a member function for retrieving the internal key extractor used.

Returns a copy of the key_from_value object used to construct the index.
Complexity: Constant.

Lookup

Hashed indices provide the full lookup functionality required by unordered associative containers, namely find, count, and equal_range. Additionally, these member functions are templatized to allow for non-standard arguments, so extending the types of search operations allowed. The kind of arguments permissible when invoking the lookup member functions is defined by the following concept.

Consider a pair (Hash, Pred) where Hash is a hash functor over values of type Key and Pred is a Binary Predicate inducing an equivalence relation on Key, whit the additional constraint that equivalent keys have the same hash value. A triplet of types (CompatibleKey, CompatibleHash, CompatiblePred) is said to be a compatible extension of (Hash, Pred) if

1. CompatibleHash is a hash functor on values of type CompatibleKey,
2. CompatiblePred is a Binary Predicate over (Key, CompatibleKey),
3. CompatiblePred is a Binary Predicate over (CompatibleKey, Key),
4. if c_eq(ck,k1) then c_eq(k1,ck),
5. if c_eq(ck,k1) and eq(k1,k2) then c_eq(ck,k2),
6. if c_eq(ck,k1) and c_eq(ck,k2) then eq(k1,k2),
7. if c_eq(ck,k1) then c_hash(ck)==hash(k1),

Additionally, a type CompatibleKey is said to be a compatible key of (Hash, Pred) if (CompatibleKey, Hash, Pred) is a compatible extension of (Hash, Pred). This implies that Hash and Pred accept arguments of type CompatibleKey, which usually means they have several overloads of therir corresponding operator() member functions.

In the context of a compatible extension or a compatible key, the expression "equivalent key" takes on its obvious interpretation.

Requires: CompatibleKey is a compatible key of (hasher, key_equal).
Effects: Returns a pointer to an element whose key is equivalent to x, or end() if such an element does not exist.
Complexity: Average case O(1) (constant), worst case O(n).

Requires: (CompatibleKey, CompatibleHash, CompatiblePred) is a compatible extension of (hasher, key_equal).
Effects: Returns a pointer to an element whose key is equivalent to x, or end() if such an element does not exist.
Complexity: Average case O(1) (constant), worst case O(n).

Requires: CompatibleKey is a compatible key of (hasher, key_equal).
Effects: Returns the number of elements with key equivalent to x.
Complexity: Average case O(count(x)), worst case O(n).

Requires: (CompatibleKey, CompatibleHash, CompatiblePred) is a compatible extension of (hasher, key_equal).
Effects: Returns the number of elements with key equivalent to x.
Complexity: Average case O(count(x,hash,eq)), worst case O(n).

Requires: CompatibleKey is a compatible key of (hasher, key_equal).
Effects: Returns a range containing all elements with keys equivalent to x (and only those).
Complexity: Average case O(count(x)), worst case O(n).

Requires: (CompatibleKey, CompatibleHash, CompatiblePred) is a compatible extension of (hasher, key_equal).
Effects: Returns a range containing all elements with keys equivalent to x (and only those).
Complexity: Average case O(count(x,hash,eq)), worst case O(n).

Hash policy

Effects: Increases if necessary the number of internal buckets so that size()/bucket_count() does not exceed the maximum load factor, and bucket_count()>=n.
Postconditions: Validity of iterators and references to the elements contained is preserved.
Complexity: Average case O(size()), worst case O(size(n)2).
Exeption safety: Strong.

Serialization

Indices cannot be serialized on their own, but only as part of the multi_index_container into which they are embedded. In describing the additional preconditions and guarantees associated to hashed indices with respect to serialization of their embedding containers, we use the concepts defined in the multi_index_container serialization section.

Requires: No additional requirements to those imposed by the container.

Requires: Additionally to the general requirements, key_eq() must be serialization-compatible with m.get<i>().key_eq(), where i is the position of the hashed index in the container.
Postconditions: On succesful loading, the range [begin(), end()) contains restored copies of every element in [m.get<i>().begin(), m.get<i>().end()), though not necessarily in the same order.

Requires: it is a valid iterator of the index. The associated multi_index_container has been previously saved.

Postconditions: On succesful loading, if it was dereferenceable then *it' is the restored copy of *it, otherwise it'==end().
Note: It is allowed that it be a const_iterator and the restored it' an iterator, or viceversa.

Requires: it is a valid local iterator of the index. The associated multi_index_container has been previously saved.

Postconditions: On succesful loading, if it was dereferenceable then *it' is the restored copy of *it; if it was m.get<i>().end(n) for some n, then it'==m'.get<i>().end(n) (where m is the original multi_index_container, m' its restored copy and i is the ordinal of the index.)
Note: It is allowed that it be a const_local_iterator and the restored it' a local_iterator, or viceversa.

Ordered indices

Boost.MultiIndex reference

Sequenced indices

Revised August 24th 2005

© Copyright 2003-2005 Joaquín M López Muñoz. Distributed under the Boost Software License, Version 1.0. (See accompanying file LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt)