summaryrefslogtreecommitdiff
path: root/stdlib/hashtbl.ml
blob: b579a6991fa25d252451517155e9aaba09935bd5 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
(**************************************************************************)
(*                                                                        *)
(*                                 OCaml                                  *)
(*                                                                        *)
(*             Xavier Leroy, projet Cristal, INRIA Rocquencourt           *)
(*                                                                        *)
(*   Copyright 1996 Institut National de Recherche en Informatique et     *)
(*     en Automatique.                                                    *)
(*                                                                        *)
(*   All rights reserved.  This file is distributed under the terms of    *)
(*   the GNU Lesser General Public License version 2.1, with the          *)
(*   special exception on linking described in the file LICENSE.          *)
(*                                                                        *)
(**************************************************************************)

(* Hash tables *)

(* We do dynamic hashing, and resize the table and rehash the elements
   when the load factor becomes too high. *)

type ('a, 'b) t =
  { mutable size: int;                        (* number of entries *)
    mutable data: ('a, 'b) bucketlist array;  (* the buckets *)
    seed: int;                        (* for randomization *)
    mutable initial_size: int;                (* initial array size *)
  }

and ('a, 'b) bucketlist =
    Empty
  | Cons of { mutable key: 'a;
              mutable data: 'b;
              mutable next: ('a, 'b) bucketlist }

(* The sign of initial_size encodes the fact that a traversal is
   ongoing or not.

   This disables the efficient in place implementation of resizing.
*)

let ongoing_traversal h =
  Obj.size (Obj.repr h) < 4 (* compatibility with old hash tables *)
  || h.initial_size < 0

let flip_ongoing_traversal h =
  h.initial_size <- - h.initial_size

(* To pick random seeds if requested *)

let randomized_default =
  let params =
    try Sys.getenv "OCAMLRUNPARAM" with Not_found ->
    try Sys.getenv "CAMLRUNPARAM" with Not_found -> "" in
  String.contains params 'R'

let randomized = Atomic.make randomized_default

let randomize () = Atomic.set randomized true
let is_randomized () = Atomic.get randomized

let prng_key = Domain.DLS.new_key Random.State.make_self_init

(* Functions which appear before the functorial interface must either be
   independent of the hash function or take it as a parameter (see #2202 and
   code below the functor definitions. *)

(* Creating a fresh, empty table *)

let rec power_2_above x n =
  if x >= n then x
  else if x * 2 > Sys.max_array_length then x
  else power_2_above (x * 2) n

let create ?(random = Atomic.get randomized) initial_size =
  let s = power_2_above 16 initial_size in
  let seed =
    if random then Random.State.bits (Domain.DLS.get prng_key) else 0
  in
  { initial_size = s; size = 0; seed = seed; data = Array.make s Empty }

let clear h =
  if h.size > 0 then begin
    h.size <- 0;
    Array.fill h.data 0 (Array.length h.data) Empty
  end

let reset h =
  let len = Array.length h.data in
  if Obj.size (Obj.repr h) < 4 (* compatibility with old hash tables *)
    || len = abs h.initial_size then
    clear h
  else begin
    h.size <- 0;
    h.data <- Array.make (abs h.initial_size) Empty
  end

let copy_bucketlist = function
  | Empty -> Empty
  | Cons {key; data; next} ->
      let rec loop prec = function
        | Empty -> ()
        | Cons {key; data; next} ->
            let r = Cons {key; data; next} in
            begin match prec with
            | Empty -> assert false
            | Cons prec ->  prec.next <- r
            end;
            loop r next
      in
      let r = Cons {key; data; next} in
      loop r next;
      r

let copy h = { h with data = Array.map copy_bucketlist h.data }

let length h = h.size

let insert_all_buckets indexfun inplace odata ndata =
  let nsize = Array.length ndata in
  let ndata_tail = Array.make nsize Empty in
  let rec insert_bucket = function
    | Empty -> ()
    | Cons {key; data; next} as cell ->
        let cell =
          if inplace then cell
          else Cons {key; data; next = Empty}
        in
        let nidx = indexfun key in
        begin match ndata_tail.(nidx) with
        | Empty -> ndata.(nidx) <- cell;
        | Cons tail -> tail.next <- cell;
        end;
        ndata_tail.(nidx) <- cell;
        insert_bucket next
  in
  for i = 0 to Array.length odata - 1 do
    insert_bucket odata.(i)
  done;
  if inplace then
    for i = 0 to nsize - 1 do
      match ndata_tail.(i) with
      | Empty -> ()
      | Cons tail -> tail.next <- Empty
    done

let resize indexfun h =
  let odata = h.data in
  let osize = Array.length odata in
  let nsize = osize * 2 in
  if nsize < Sys.max_array_length then begin
    let ndata = Array.make nsize Empty in
    let inplace = not (ongoing_traversal h) in
    h.data <- ndata;          (* so that indexfun sees the new bucket count *)
    insert_all_buckets (indexfun h) inplace odata ndata
  end

let iter f h =
  let rec do_bucket = function
    | Empty ->
        ()
    | Cons{key; data; next} ->
        f key data; do_bucket next in
  let old_trav = ongoing_traversal h in
  if not old_trav then flip_ongoing_traversal h;
  try
    let d = h.data in
    for i = 0 to Array.length d - 1 do
      do_bucket d.(i)
    done;
    if not old_trav then flip_ongoing_traversal h;
  with exn when not old_trav ->
    flip_ongoing_traversal h;
    raise exn

let rec filter_map_inplace_bucket f h i prec = function
  | Empty ->
      begin match prec with
      | Empty -> h.data.(i) <- Empty
      | Cons c -> c.next <- Empty
      end
  | (Cons ({key; data; next} as c)) as slot ->
      begin match f key data with
      | None ->
          h.size <- h.size - 1;
          filter_map_inplace_bucket f h i prec next
      | Some data ->
          begin match prec with
          | Empty -> h.data.(i) <- slot
          | Cons c -> c.next <- slot
          end;
          c.data <- data;
          filter_map_inplace_bucket f h i slot next
      end

let filter_map_inplace f h =
  let d = h.data in
  let old_trav = ongoing_traversal h in
  if not old_trav then flip_ongoing_traversal h;
  try
    for i = 0 to Array.length d - 1 do
      filter_map_inplace_bucket f h i Empty h.data.(i)
    done;
    if not old_trav then flip_ongoing_traversal h
  with exn when not old_trav ->
    flip_ongoing_traversal h;
    raise exn

let fold f h init =
  let rec do_bucket b accu =
    match b with
      Empty ->
        accu
    | Cons{key; data; next} ->
        do_bucket next (f key data accu) in
  let old_trav = ongoing_traversal h in
  if not old_trav then flip_ongoing_traversal h;
  try
    let d = h.data in
    let accu = ref init in
    for i = 0 to Array.length d - 1 do
      accu := do_bucket d.(i) !accu
    done;
    if not old_trav then flip_ongoing_traversal h;
    !accu
  with exn when not old_trav ->
    flip_ongoing_traversal h;
    raise exn

type statistics = {
  num_bindings: int;
  num_buckets: int;
  max_bucket_length: int;
  bucket_histogram: int array
}

let rec bucket_length accu = function
  | Empty -> accu
  | Cons{next} -> bucket_length (accu + 1) next

let stats h =
  let mbl =
    Array.fold_left (fun m b -> Int.max m (bucket_length 0 b)) 0 h.data in
  let histo = Array.make (mbl + 1) 0 in
  Array.iter
    (fun b ->
      let l = bucket_length 0 b in
      histo.(l) <- histo.(l) + 1)
    h.data;
  { num_bindings = h.size;
    num_buckets = Array.length h.data;
    max_bucket_length = mbl;
    bucket_histogram = histo }

(** {1 Iterators} *)

let to_seq tbl =
  (* capture current array, so that even if the table is resized we
     keep iterating on the same array *)
  let tbl_data = tbl.data in
  (* state: index * next bucket to traverse *)
  let rec aux i buck () = match buck with
    | Empty ->
        if i = Array.length tbl_data
        then Seq.Nil
        else aux(i+1) tbl_data.(i) ()
    | Cons {key; data; next} ->
        Seq.Cons ((key, data), aux i next)
  in
  aux 0 Empty

let to_seq_keys m = Seq.map fst (to_seq m)

let to_seq_values m = Seq.map snd (to_seq m)

(* Functorial interface *)

module type HashedType =
  sig
    type t
    val equal: t -> t -> bool
    val hash: t -> int
  end

module type SeededHashedType =
  sig
    type t
    val equal: t -> t -> bool
    val seeded_hash: int -> t -> int
  end

module type S =
  sig
    type key
    type !'a t
    val create: int -> 'a t
    val clear : 'a t -> unit
    val reset : 'a t -> unit
    val copy: 'a t -> 'a t
    val add: 'a t -> key -> 'a -> unit
    val remove: 'a t -> key -> unit
    val find: 'a t -> key -> 'a
    val find_opt: 'a t -> key -> 'a option
    val find_all: 'a t -> key -> 'a list
    val replace : 'a t -> key -> 'a -> unit
    val mem : 'a t -> key -> bool
    val iter: (key -> 'a -> unit) -> 'a t -> unit
    val filter_map_inplace: (key -> 'a -> 'a option) -> 'a t -> unit
    val fold: (key -> 'a -> 'b -> 'b) -> 'a t -> 'b -> 'b
    val length: 'a t -> int
    val stats: 'a t -> statistics
    val to_seq : 'a t -> (key * 'a) Seq.t
    val to_seq_keys : _ t -> key Seq.t
    val to_seq_values : 'a t -> 'a Seq.t
    val add_seq : 'a t -> (key * 'a) Seq.t -> unit
    val replace_seq : 'a t -> (key * 'a) Seq.t -> unit
    val of_seq : (key * 'a) Seq.t -> 'a t
  end

module type SeededS =
  sig
    type key
    type !'a t
    val create : ?random:bool -> int -> 'a t
    val clear : 'a t -> unit
    val reset : 'a t -> unit
    val copy : 'a t -> 'a t
    val add : 'a t -> key -> 'a -> unit
    val remove : 'a t -> key -> unit
    val find : 'a t -> key -> 'a
    val find_opt: 'a t -> key -> 'a option
    val find_all : 'a t -> key -> 'a list
    val replace : 'a t -> key -> 'a -> unit
    val mem : 'a t -> key -> bool
    val iter : (key -> 'a -> unit) -> 'a t -> unit
    val filter_map_inplace: (key -> 'a -> 'a option) -> 'a t -> unit
    val fold : (key -> 'a -> 'b -> 'b) -> 'a t -> 'b -> 'b
    val length : 'a t -> int
    val stats: 'a t -> statistics
    val to_seq : 'a t -> (key * 'a) Seq.t
    val to_seq_keys : _ t -> key Seq.t
    val to_seq_values : 'a t -> 'a Seq.t
    val add_seq : 'a t -> (key * 'a) Seq.t -> unit
    val replace_seq : 'a t -> (key * 'a) Seq.t -> unit
    val of_seq : (key * 'a) Seq.t -> 'a t
  end

module MakeSeeded(H: SeededHashedType): (SeededS with type key = H.t) =
  struct
    type key = H.t
    type 'a hashtbl = (key, 'a) t
    type 'a t = 'a hashtbl
    let create = create
    let clear = clear
    let reset = reset
    let copy = copy

    let key_index h key =
      (H.seeded_hash h.seed key) land (Array.length h.data - 1)

    let add h key data =
      let i = key_index h key in
      let bucket = Cons{key; data; next=h.data.(i)} in
      h.data.(i) <- bucket;
      h.size <- h.size + 1;
      if h.size > Array.length h.data lsl 1 then resize key_index h

    let rec remove_bucket h i key prec = function
      | Empty ->
          ()
      | (Cons {key=k; next}) as c ->
          if H.equal k key
          then begin
            h.size <- h.size - 1;
            match prec with
            | Empty -> h.data.(i) <- next
            | Cons c -> c.next <- next
          end
          else remove_bucket h i key c next

    let remove h key =
      let i = key_index h key in
      remove_bucket h i key Empty h.data.(i)

    let rec find_rec key = function
      | Empty ->
          raise Not_found
      | Cons{key=k; data; next} ->
          if H.equal key k then data else find_rec key next

    let find h key =
      match h.data.(key_index h key) with
      | Empty -> raise Not_found
      | Cons{key=k1; data=d1; next=next1} ->
          if H.equal key k1 then d1 else
          match next1 with
          | Empty -> raise Not_found
          | Cons{key=k2; data=d2; next=next2} ->
              if H.equal key k2 then d2 else
              match next2 with
              | Empty -> raise Not_found
              | Cons{key=k3; data=d3; next=next3} ->
                  if H.equal key k3 then d3 else find_rec key next3

    let rec find_rec_opt key = function
      | Empty ->
          None
      | Cons{key=k; data; next} ->
          if H.equal key k then Some data else find_rec_opt key next

    let find_opt h key =
      match h.data.(key_index h key) with
      | Empty -> None
      | Cons{key=k1; data=d1; next=next1} ->
          if H.equal key k1 then Some d1 else
          match next1 with
          | Empty -> None
          | Cons{key=k2; data=d2; next=next2} ->
              if H.equal key k2 then Some d2 else
              match next2 with
              | Empty -> None
              | Cons{key=k3; data=d3; next=next3} ->
                  if H.equal key k3 then Some d3 else find_rec_opt key next3

    let find_all h key =
      let[@tail_mod_cons] rec find_in_bucket = function
      | Empty ->
          []
      | Cons{key=k; data=d; next} ->
          if H.equal k key
          then d :: find_in_bucket next
          else find_in_bucket next in
      find_in_bucket h.data.(key_index h key)

    let rec replace_bucket key data = function
      | Empty ->
          true
      | Cons ({key=k; next} as slot) ->
          if H.equal k key
          then (slot.key <- key; slot.data <- data; false)
          else replace_bucket key data next

    let replace h key data =
      let i = key_index h key in
      let l = h.data.(i) in
      if replace_bucket key data l then begin
        h.data.(i) <- Cons{key; data; next=l};
        h.size <- h.size + 1;
        if h.size > Array.length h.data lsl 1 then resize key_index h
      end

    let rec mem_in_bucket key = function
      | Empty ->
          false
      | Cons{key=k; next} ->
          H.equal k key || mem_in_bucket key next

    let mem h key =
      mem_in_bucket key h.data.(key_index h key)

    let add_seq tbl i =
      Seq.iter (fun (k,v) -> add tbl k v) i

    let replace_seq tbl i =
      Seq.iter (fun (k,v) -> replace tbl k v) i

    let of_seq i =
      let tbl = create 16 in
      replace_seq tbl i;
      tbl

    let iter = iter
    let filter_map_inplace = filter_map_inplace
    let fold = fold
    let length = length
    let stats = stats
    let to_seq = to_seq
    let to_seq_keys = to_seq_keys
    let to_seq_values = to_seq_values
  end

module Make(H: HashedType): (S with type key = H.t) =
  struct
    include MakeSeeded(struct
        type t = H.t
        let equal = H.equal
        let seeded_hash (_seed: int) x = H.hash x
      end)
    let create sz = create ~random:false sz
    let of_seq i =
      let tbl = create 16 in
      replace_seq tbl i;
      tbl
  end

(* Polymorphic hash function-based tables *)
(* Code included below the functorial interface to guard against accidental
   use - see #2202 *)

external seeded_hash_param :
  int -> int -> int -> 'a -> int = "caml_hash" [@@noalloc]

let hash x = seeded_hash_param 10 100 0 x
let hash_param n1 n2 x = seeded_hash_param n1 n2 0 x
let seeded_hash seed x = seeded_hash_param 10 100 seed x

let key_index h key =
  if Obj.size (Obj.repr h) >= 4
  then (seeded_hash_param 10 100 h.seed key) land (Array.length h.data - 1)
  else invalid_arg "Hashtbl: unsupported hash table format"

let add h key data =
  let i = key_index h key in
  let bucket = Cons{key; data; next=h.data.(i)} in
  h.data.(i) <- bucket;
  h.size <- h.size + 1;
  if h.size > Array.length h.data lsl 1 then resize key_index h

let rec remove_bucket h i key prec = function
  | Empty ->
      ()
  | (Cons {key=k; next}) as c ->
      if compare k key = 0
      then begin
        h.size <- h.size - 1;
        match prec with
        | Empty -> h.data.(i) <- next
        | Cons c -> c.next <- next
      end
      else remove_bucket h i key c next

let remove h key =
  let i = key_index h key in
  remove_bucket h i key Empty h.data.(i)

let rec find_rec key = function
  | Empty ->
      raise Not_found
  | Cons{key=k; data; next} ->
      if compare key k = 0 then data else find_rec key next

let find h key =
  match h.data.(key_index h key) with
  | Empty -> raise Not_found
  | Cons{key=k1; data=d1; next=next1} ->
      if compare key k1 = 0 then d1 else
      match next1 with
      | Empty -> raise Not_found
      | Cons{key=k2; data=d2; next=next2} ->
          if compare key k2 = 0 then d2 else
          match next2 with
          | Empty -> raise Not_found
          | Cons{key=k3; data=d3; next=next3} ->
              if compare key k3 = 0 then d3 else find_rec key next3

let rec find_rec_opt key = function
  | Empty ->
      None
  | Cons{key=k; data; next} ->
      if compare key k = 0 then Some data else find_rec_opt key next

let find_opt h key =
  match h.data.(key_index h key) with
  | Empty -> None
  | Cons{key=k1; data=d1; next=next1} ->
      if compare key k1 = 0 then Some d1 else
      match next1 with
      | Empty -> None
      | Cons{key=k2; data=d2; next=next2} ->
          if compare key k2 = 0 then Some d2 else
          match next2 with
          | Empty -> None
          | Cons{key=k3; data=d3; next=next3} ->
              if compare key k3 = 0 then Some d3 else find_rec_opt key next3

let find_all h key =
  let[@tail_mod_cons] rec find_in_bucket = function
  | Empty ->
      []
  | Cons{key=k; data; next} ->
      if compare k key = 0
      then data :: find_in_bucket next
      else find_in_bucket next in
  find_in_bucket h.data.(key_index h key)

let rec replace_bucket key data = function
  | Empty ->
      true
  | Cons ({key=k; next} as slot) ->
      if compare k key = 0
      then (slot.key <- key; slot.data <- data; false)
      else replace_bucket key data next

let replace h key data =
  let i = key_index h key in
  let l = h.data.(i) in
  if replace_bucket key data l then begin
    h.data.(i) <- Cons{key; data; next=l};
    h.size <- h.size + 1;
    if h.size > Array.length h.data lsl 1 then resize key_index h
  end

let rec mem_in_bucket key = function
  | Empty ->
      false
  | Cons{key=k; next} ->
      compare k key = 0 || mem_in_bucket key next

let mem h key =
  mem_in_bucket key h.data.(key_index h key)

let add_seq tbl i =
  Seq.iter (fun (k,v) -> add tbl k v) i

let replace_seq tbl i =
  Seq.iter (fun (k,v) -> replace tbl k v) i

let of_seq i =
  let tbl = create 16 in
  replace_seq tbl i;
  tbl

let rebuild ?(random = Atomic.get randomized) h =
  let s = power_2_above 16 (Array.length h.data) in
  let seed =
    if random then Random.State.bits (Domain.DLS.get prng_key)
    else if Obj.size (Obj.repr h) >= 4 then h.seed
    else 0 in
  let h' = {
    size = h.size;
    data = Array.make s Empty;
    seed = seed;
    initial_size = if Obj.size (Obj.repr h) >= 4 then h.initial_size else s
  } in
  insert_all_buckets (key_index h') false h.data h'.data;
  h'