From 8c06c13563ba27b64c9cf79db9d6875bef253a2e Mon Sep 17 00:00:00 2001
From: Simon Cruanes <simon.cruanes.2007@m4x.org>
Date: Sat, 22 Aug 2015 23:49:36 +0200
Subject: [PATCH] add `group_by` (ignores the ordering)

---
 sequence.ml            | 16 ++++++++++++++++
 sequence.mli           |  5 +++++
 tests/test_sequence.ml | 12 +++++++++---
 3 files changed, 30 insertions(+), 3 deletions(-)

diff --git a/sequence.ml b/sequence.ml
index 1401c8d..c781d53 100644
--- a/sequence.ml
+++ b/sequence.ml
@@ -253,6 +253,22 @@ let group_succ_by ?(eq=fun x y -> x = y) seq k =
 
 let group = group_succ_by
 
+let group_by (type k) ?(hash=Hashtbl.hash) ?(eq=(=)) seq =
+  let module Tbl = Hashtbl.Make(struct
+    type t = k
+    let equal = eq
+    let hash = hash
+  end) in
+  (* compute group table *)
+  let tbl = Tbl.create 32 in
+  seq
+    (fun x ->
+      let l = try Tbl.find tbl x with Not_found -> [] in
+      Tbl.replace tbl x (x::l)
+    );
+  fun yield ->
+    Tbl.iter (fun _ l -> yield l) tbl
+
 let uniq ?(eq=fun x y -> x = y) seq k =
   let has_prev = ref false
   and prev = ref (Obj.magic 0) in  (* avoid option type, costly *)
diff --git a/sequence.mli b/sequence.mli
index 4104b02..56969c5 100644
--- a/sequence.mli
+++ b/sequence.mli
@@ -223,6 +223,11 @@ val group_succ_by : ?eq:('a -> 'a -> bool) -> 'a t -> 'a list t
       Synonym to {!group}.
       @since NEXT_RELEASE *)
 
+val group_by : ?hash:('a -> int) -> ?eq:('a -> 'a -> bool) ->
+               'a t -> 'a list t
+  (** Group equal elements, disregarding their order of appearance.
+      The result sequence is traversable as many times as required.
+      @since NEXT_RELEASE *)
 
 val uniq : ?eq:('a -> 'a -> bool) -> 'a t -> 'a t
   (** Remove consecutive duplicate elements. Basically this is
diff --git a/tests/test_sequence.ml b/tests/test_sequence.ml
index 8fd1b17..a85a5c6 100644
--- a/tests/test_sequence.ml
+++ b/tests/test_sequence.ml
@@ -124,11 +124,16 @@ let test_sort_uniq () =
     |> S.to_list
     |> OUnit.assert_equal [1;2;3;4;5;42]
 
-let test_group () =
+let test_group_succ () =
   [1;2;3;3;2;2;3;4]
-    |> S.of_list |> S.group ?eq:None |> S.to_list
+    |> S.of_list |> S.group_succ_by ?eq:None |> S.to_list
     |> OUnit.assert_equal [[1];[2];[3;3];[2;2];[3];[4]]
 
+let test_group_by () =
+  [1;2;3;3;2;2;3;4]
+    |> S.of_list |> S.group_by ?eq:None ?hash:None |> S.sort ?cmp:None |> S.to_list
+    |> OUnit.assert_equal [[1];[2;2;2];[3;3;3];[4]]
+
 let test_uniq () =
   [1;2;2;3;4;4;4;3;3]
     |> S.of_list |> S.uniq ?eq:None |> S.to_list
@@ -226,7 +231,8 @@ let suite =
       "test_big_persistent" >:: test_big_persistent;
       "test_sort" >:: test_sort;
       "test_sort_uniq" >:: test_sort;
-      "test_group" >:: test_group;
+      "test_group_succ_by" >:: test_group_succ;
+      "test_group_by" >:: test_group_by;
       "test_uniq" >:: test_uniq;
       "test_product" >:: test_product;
       "test_join" >:: test_join;