This module implements applicative association tables, also known as finite maps or dictionaries, given a total ordering function over the keys. All operations over maps are purely applicative (no side-effects). The implementation uses balanced binary trees, and therefore searching and insertion take time logarithmic in the size of the map.
Note OCaml, Batteries Included, provides two implementations of maps: polymorphic maps and functorized maps. Functorized maps (see S and Make) are slightly more complex to use but offer stronger type-safety. Polymorphic maps make it easier to shoot yourself in the foot. In case of doubt, you should use functorized maps.
Functorized maps
The important part is the Make module which builds association maps from a user-provided datatype and comparison function. In the Make module (or its output signature S) are documentated all functions available on maps.
Here is a typical example of use:
module MyKeyType = struct
type t = my_type
let compare = my_compare_function
end
module MyMap = Map.Make(MyKeyType)
let some_map = MyMap.add something MyMap.empty
...
The functions below present the manipulation of polymorphic maps, as were provided by the Extlib PMap module.
They are similar in functionality to the functorized Make module, but only uses the Pervasives.compare function to compare elements. If you need to compare using a custom comparison function, it is recommended to use the functorized maps provided by Make.
add x y m returns a map containing the same bindings as m, plus a binding of x to y. If x was already bound in m, its previous binding disappears. If x was already bound to some z that is physically equal to y, then the returned map is physically equal to m.
update k1 k2 v2 m replace the previous binding of k1 in m by k2 associated to v2. This is equivalent to add k2 v2 (remove k1) m, but more efficient in the case where k1 and k2 have the same key ordering. If k1 and k2 have the same key ordering and v2 is physically equal to the value k1 is bound to in m then the returned map will be physically equal to m
update_stdlib k f m returns a map containing the same bindings as m, except k has a new binding as determined by f: First, calculate y as f (find_opt k m). If y = Some v then k will be bound to v in the resulting map. Else k will not be bound in the resulting map.
If v is physically equal to the value of the previous binding of k in m, then the returned map will be physically equal to m.
This function does the same thing as update in the stdlib, but has a different name for backwards compatibility reasons.
find_first f m returns the first binding (k, v) for which f k is true or raises Not_found if there is no such binding. f must be monotonically increasing, i.e. if k1 < k2 && f k1 is true then f k2 must also be true.
find_first_opt f m returns Some (k, v) for the first binding (k, v) for which f k is true or returns None if there is no such binding. f must be monotonically increasing, i.e. if k1 < k2 && f k1 is true then f k2 must also be true.
find_last f m returns the last binding (k, v) for which f k is true or raises Not_found if there is no such binding. f must be monotonically decreasing, i.e. if k1 < k2 && f k2 is true then f k1 must also be true.
find_last_opt f m returns Some (k, v) for the last binding (k, v) for which f k is true or returns None if there is no such binding. f must be monotonically decreasing, i.e. if k1 < k2 && f k2 is true then f k1 must also be true.
remove x m returns a map containing the same bindings as m, except for x which is unbound in the returned map. The returned map is physically equal to the passed one if x was already unbound.
mem x m returns true if m contains a binding for x, and false otherwise.
Sourceval iter : ('a->'b-> unit)->('a, 'b)t-> unit
iter f m applies f to all bindings in map m. f receives the key as first argument, and the associated value as second argument. The order in which the bindings are passed to f is unspecified. Only current bindings are presented to f: bindings hidden by more recent bindings are not passed to f.
map f m returns a map with same domain as m, where the associated value a of all bindings of m has been replaced by the result of the application of f to a. The order in which the associated values are passed to f is unspecified.
fold f m a computes (f kN dN ... (f k1 d1 (f k0 d0 a))...), where k0,k1..kN are the keys of all bindings in m, and d0,d1..dN are the associated data. The order in which the bindings are presented to f is unspecified.
Same as fold, but the function receives as arguments both the key and the associated value for each binding of the map.
Sourceval at_rank_exn : int ->('key, 'a)t->'key * 'a
at_rank_exn i m returns the (key,value) pair whose key is at rank i in m, that is the i-th element in increasing order of the keys (the 0-th element being the smallest key in m with its associated value).
filterv f m returns a map where only the values a of m such that f a = true remain. The bindings are passed to f in increasing order with respect to the ordering over the type of the keys.
filter f m returns a map where only the (key, value) pairs of m such that f key value = true remain. The bindings are passed to f in increasing order with respect to the ordering over the type of the keys. If f returns true for all bindings of m the returned map is physically equal to m.
filter_map f m combines the features of filter and map. It calls calls f key0 a0, f key1 a1, f keyn an where a0..an are the elements of m and key0..keyn the respective corresponding keys. It returns the map of (keyi, bi) pairs such as f keyi ai = Some bi (when f returns None, the corresponding element of m is discarded).
Return Some (k, v) for one binding (k, v) of the given map, if the map is not empty. Else, return None. Which binding is chosen is unspecified, but equal bindings will be chosen for equal maps.
Return one binding of the given map. The difference with choose is that there is no guarantee that equals elements will be picked for equal sets. This merely returns the quickest binding to get (O(1)).
split x m returns a triple (l, data, r), where l is the map with all the bindings of m whose key is strictly less than x; r is the map with all the bindings of m whose key is strictly greater than x; data is None if m contains no binding for x, or Some v if m binds v to x.
partition p m returns a pair of maps (m1, m2), where m1 contains all the bindings of s that satisfy the predicate p, and m2 is the map with all the bindings of s that do not satisfy p.
modify k f m replaces the previous binding for k with f applied to that value. If k is unbound in m or Not_found is raised during the search, Not_found is raised.
modify_def v0 k f m replaces the previous binding for k with f applied to that value. If k is unbound in m or Not_found is raised during the search, f v0 is inserted (as if the value found were v0).
union m1 m2 merges two maps, using the comparison function of m1. In case of conflicted bindings, m2's bindings override m1's. Equivalent to foldi add m2 m1. The resulting map uses the comparison function of m1.
union_stdlib f m1 m2 computes a map whose keys are a subset of the keys of m1 and of m2. When the same binding is defined in both arguments, the function f is used to combine them. This function is similar to merge, except f is only called if a key is present in both m1 and m2. If a key is present in either m1 or m2 but not in both, it (and the corresponding value) will be present in the resulting map.
This is the union method from the stdlib map, renamed for backwards compatibility.
diff m1 m2 removes all bindings of keys found in m2 from m1, using the comparison function of m1. Equivalent to foldi (fun k _v m -> remove k m) m2 m1. The resulting map uses the comparison function of m1.
intersect merge_f m1 m2 returns a map with bindings only for keys bound in both m1 and m2, and with k bound to merge_f v1 v2, where v1 and v2 are k's bindings in m1 and m2. The resulting map uses the comparison function of m1.
merge f m1 m2 computes a map whose keys is a subset of keys of m1 and of m2. The presence of each such binding, and the corresponding value, is determined with the function f. The resulting map uses the comparison function of m1.
Sourceval compare : ('b->'b-> int)->('a, 'b)t->('a, 'b)t-> int
Construct a comparison or equality function for maps based on a value comparison or equality function. Uses the key comparison function to compare keys