1 /* PSPP - a program for statistical analysis.
2 Copyright (C) 2008 Free Software Foundation, Inc.
4 This program is free software: you can redistribute it and/or modify
5 it under the terms of the GNU General Public License as published by
6 the Free Software Foundation, either version 3 of the License, or
7 (at your option) any later version.
9 This program is distributed in the hope that it will be useful,
10 but WITHOUT ANY WARRANTY; without even the implied warranty of
11 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
12 GNU General Public License for more details.
14 You should have received a copy of the GNU General Public License
15 along with this program. If not, see <http://www.gnu.org/licenses/>. */
17 /* Hash table with separate chaining.
19 This header (hmap.h) supplies an "embedded" implementation of
20 a hash table that uses linked lists to resolve collisions
21 ("separate chaining"). Its companion header (hmapx.h)
22 supplies a "external" implementation that is otherwise
23 similar. The two variants are described briefly here. The
24 embedded variant, for which this is the header, is described
25 in slightly more detail below. Each function also has a
26 detailed usage comment at its point of definition. (Many of
27 those definitions are inline in this file, because they are so
28 simple. Others are in hmap.c.)
30 The "hmap" embedded hash table implementation puts the hash
31 table node (which includes the linked list used for resolving
32 collisions) within the data structure that the hash table
33 contains. This makes allocation efficient, in space and time,
34 because no additional call into an allocator is needed to
35 obtain memory for the hash table node. It also makes it easy
36 to find the hash table node associated with a given object.
37 However, it's difficult to include a given object in an
38 arbitrary number of hash tables.
40 The "hmapx" external hash table implementation allocates hash
41 table nodes separately from the objects in the hash table.
42 Inserting and removing hash table elements requires dynamic
43 allocation, so it is normally slower and takes more memory
44 than the embedded implementation. It also requires searching
45 the table to find the node associated with a given object.
46 However, it's easy to include a given object in an arbitrary
47 number of hash tables. It's also possible to create an
48 external hash table without adding a member to the data
49 structure that the hash table contains. */
51 #ifndef LIBPSPP_HMAP_H
52 #define LIBPSPP_HMAP_H 1
54 /* Embedded hash table with separate chaining.
56 To create an embedded hash table, declare an instance of
57 struct hmap, then initialize it with hmap_init():
61 struct hmap map = HMAP_INITIALIZER (map);
63 Each node in the hash table, presumably a structure type, must
64 include a struct hmap_node member. Here's an example:
67 struct hmap_node node; // hmap_node member.
68 const char *string; // Another member.
70 The hash table functions work with pointers to struct
71 hmap_node. To obtain a pointer to your structure type given a
72 pointer to struct hmap_node, use the HMAP_DATA macro.
74 Inserting and deleting elements is straightforward. Use
75 hmap_insert() to insert an element and hmap_delete() to delete
78 my_foo.string = "My string";
79 hmap_insert (&map, &my_foo.node, hsh_hash_string (my_foo.string));
81 hmap_delete (&map, &my_foo.node);
82 You must pass the element's hash value as one of
83 hmap_insert()'s arguments. The hash table saves this hash
84 value for use later to speed searches and to rehash as the
87 hmap_insert() does not check whether the newly inserted
88 element duplicates an element already in the hash table. The
89 client is responsible for doing so, if this is desirable.
91 The hash table does not provide a direct way to search for an
92 existing element. Instead, it provides the means to iterate
93 over all the elements in the hash table with a given hash
94 value. It is easy to compose a search function from such a
95 building block. For example:
97 find_foo (const struct hmap *map, const char *name)
99 const struct foo *foo;
102 hash = hsh_hash_string (name);
103 HMAP_FOR_EACH_WITH_HASH (foo, struct foo, node, hash, map)
104 if (!strcmp (foo->name, name))
109 Here is how to iterate through the elements currently in the
112 HMAP_FOR_EACH (foo, struct foo, node, &map)
114 ...do something with foo...
120 /* Returns the data structure corresponding to the given NODE,
121 assuming that NODE is embedded as the given MEMBER name in
122 data type STRUCT. NODE must not be a null pointer. */
123 #define HMAP_DATA(NODE, STRUCT, MEMBER) \
124 ((STRUCT *) ((char *) (NODE) - offsetof (STRUCT, MEMBER)))
126 /* Like HMAP_DATA, except that a null NODE yields a null pointer
128 #define HMAP_NULLABLE_DATA(NODE, STRUCT, MEMBER) \
129 hmap_nullable_data__ (NODE, offsetof (STRUCT, MEMBER))
131 /* Hash table node. */
134 struct hmap_node *next; /* Next in chain. */
135 size_t hash; /* Hash value. */
138 static inline size_t hmap_node_hash (const struct hmap_node *);
143 size_t count; /* Number of inserted nodes. */
144 size_t mask; /* Number of buckets (power of 2), minus 1. */
145 struct hmap_node **buckets; /* Array of buckets. */
146 struct hmap_node *one; /* One bucket, to eliminate corner cases. */
149 /* Suitable for use as the initializer for a struct hmap named
151 struct hmap map = HMAP_INITIALIZER (map);
152 HMAP_INITIALIZER() is an alternative to hmap_init(). */
153 #define HMAP_INITIALIZER(MAP) { 0, 0, &(MAP).one, NULL }
155 /* Creation and destruction. */
156 void hmap_init (struct hmap *);
157 void hmap_swap (struct hmap *, struct hmap *);
158 void hmap_destroy (struct hmap *);
160 /* Storage management. */
161 void hmap_reserve (struct hmap *, size_t capacity);
162 void hmap_shrink (struct hmap *);
164 /* Search. Refer to the large comment near the top of this file
166 static inline struct hmap_node *hmap_first_with_hash (const struct hmap *,
168 static inline struct hmap_node *hmap_next_with_hash (const struct hmap_node *);
170 /* Insertion and deletion. */
171 static inline void hmap_insert (struct hmap *, struct hmap_node *,
173 static inline void hmap_insert_fast (struct hmap *, struct hmap_node *,
175 static inline void hmap_delete (struct hmap *, struct hmap_node *);
178 static inline struct hmap_node *hmap_first (const struct hmap *);
179 static inline struct hmap_node *hmap_next (const struct hmap *,
180 const struct hmap_node *);
183 static inline size_t hmap_count (const struct hmap *);
184 static inline size_t hmap_capacity (const struct hmap *);
186 /* Updating data elements. */
187 void hmap_changed (struct hmap *, struct hmap_node *, size_t new_hash);
188 void hmap_moved (struct hmap *,
189 struct hmap_node *, const struct hmap_node *old);
191 /* Convenience macros for search.
193 These macros automatically use HMAP_DATA to obtain the data
194 elements that encapsulate hmap nodes, which often saves typing
195 and can make code easier to read. Refer to the large comment
196 near the top of this file for an example.
198 These macros evaluate HASH only once. They evaluate their
199 other arguments many times. */
200 #define HMAP_FIRST_WITH_HASH(STRUCT, MEMBER, HMAP, HASH) \
201 HMAP_NULLABLE_DATA (hmap_first_with_hash (HMAP, HASH), STRUCT, MEMBER)
202 #define HMAP_NEXT_WITH_HASH(DATA, STRUCT, MEMBER) \
203 HMAP_NULLABLE_DATA (hmap_next_with_hash (&(DATA)->MEMBER), STRUCT, MEMBER)
204 #define HMAP_FOR_EACH_WITH_HASH(DATA, STRUCT, MEMBER, HASH, HMAP) \
205 for ((DATA) = HMAP_FIRST_WITH_HASH (STRUCT, MEMBER, HMAP, HASH); \
207 (DATA) = HMAP_NEXT_WITH_HASH (DATA, STRUCT, MEMBER))
208 #define HMAP_FOR_EACH_WITH_HASH_SAFE(DATA, NEXT, STRUCT, MEMBER, HASH, HMAP) \
209 for ((DATA) = HMAP_FIRST_WITH_HASH (STRUCT, MEMBER, HMAP, HASH); \
211 ? ((NEXT) = HMAP_NEXT_WITH_HASH (DATA, STRUCT, MEMBER), 1) \
215 /* Convenience macros for iteration.
217 These macros automatically use HMAP_DATA to obtain the data
218 elements that encapsulate hmap nodes, which often saves typing
219 and can make code easier to read. Refer to the large comment
220 near the top of this file for an example.
222 These macros evaluate their arguments many times. */
223 #define HMAP_FIRST(STRUCT, MEMBER, HMAP) \
224 HMAP_NULLABLE_DATA (hmap_first (HMAP), STRUCT, MEMBER)
225 #define HMAP_NEXT(DATA, STRUCT, MEMBER, HMAP) \
226 HMAP_NULLABLE_DATA (hmap_next (HMAP, &(DATA)->MEMBER), STRUCT, MEMBER)
227 #define HMAP_FOR_EACH(DATA, STRUCT, MEMBER, HMAP) \
228 for ((DATA) = HMAP_FIRST (STRUCT, MEMBER, HMAP); \
230 (DATA) = HMAP_NEXT (DATA, STRUCT, MEMBER, HMAP))
231 #define HMAP_FOR_EACH_SAFE(DATA, NEXT, STRUCT, MEMBER, HMAP) \
232 for ((DATA) = HMAP_FIRST (STRUCT, MEMBER, HMAP); \
234 ? ((NEXT) = HMAP_NEXT (DATA, STRUCT, MEMBER, HMAP), 1) \
238 /* Inline definitions. */
240 static inline struct hmap_node *hmap_find_hash__ (struct hmap_node *, size_t);
241 static inline struct hmap_node *hmap_first_nonempty_bucket__ (
242 const struct hmap *, size_t start);
243 static inline size_t hmap_mask_to_capacity__ (size_t mask);
245 /* Returns the hash value associated with NODE. */
247 hmap_node_hash (const struct hmap_node *node)
252 /* Returns the first node in MAP that has hash value HASH, or a
253 null pointer if MAP does not contain any node with that hash
256 Assuming uniform hashing and no duplicate data items in MAP,
257 this function runs in constant time. (Amortized over an
258 iteration over all data items with a given HASH, its runtime
259 is proportional to the length of the hash chain for HASH, so
260 given a pathological hash function, e.g. one that returns a
261 constant value, its runtime degenerates to linear in the
262 length of NODE's hash chain.)
264 Nodes are returned in arbitrary order that may change whenever
265 the hash table's current capacity changes, as reported by
266 hmap_capacity(). Calls to hmap_insert(), hmap_reserve(), and
267 hmap_shrink() can change the capacity of a hash map.
268 Inserting a node with hmap_insert_fast() or deleting one with
269 hmap_delete() will not change the relative ordering of nodes.
271 The HMAP_FOR_EACH_WITH_HASH and HMAP_FOR_EACH_WITH_HASH_SAFE
272 macros provide convenient ways to iterate over all the nodes
273 with a given hash. The HMAP_FIRST_WITH_HASH macro is an
274 interface to this particular function that is often more
276 static inline struct hmap_node *
277 hmap_first_with_hash (const struct hmap *map, size_t hash)
279 return hmap_find_hash__ (map->buckets[hash & map->mask], hash);
282 /* Returns the next node in MAP after NODE that has the same hash
283 value as NODE, or a null pointer if MAP does not contain any
284 more nodes with that hash value.
286 Assuming uniform hashing and no duplicate data items in MAP,
287 this function runs in constant time. (Amortized over an
288 iteration over all data items with a given HASH, its runtime
289 is proportional to the length of the hash chain for HASH, so
290 given a pathological hash function, e.g. one that returns a
291 constant value, its runtime degenerates to linear in the
292 length of NODE's hash chain.)
294 Nodes are returned in arbitrary order that may change whenever
295 the hash table's current capacity changes, as reported by
296 hmap_capacity(). Calls to hmap_insert(), hmap_reserve(), and
297 hmap_shrink() can change the capacity of a hash map.
298 Inserting a node with hmap_insert_fast() or deleting one with
299 hmap_delete() will not change the relative ordering of nodes.
301 The HMAP_FOR_EACH_WITH_HASH and HMAP_FOR_EACH_WITH_HASH_SAFE
302 macros provide convenient ways to iterate over all the nodes
303 with a given hash. The HMAP_NEXT_WITH_HASH macro is an
304 interface to this particular function that is often more
306 static inline struct hmap_node *
307 hmap_next_with_hash (const struct hmap_node *node)
309 return hmap_find_hash__ (node->next, node->hash);
312 /* Inserts NODE into MAP with hash value HASH. If the insertion
313 causes MAP's current capacity, as reported by hmap_capacity(),
314 to be exceeded, rehashes MAP with an increased number of hash
317 This function runs in constant time amortized over all the
320 This function does not verify that MAP does not already
321 contain a data item with the same value as NODE. If
322 duplicates should be disallowed (which is the usual case),
323 then the client must check for duplicates itself before
324 inserting the new node. */
326 hmap_insert (struct hmap *map, struct hmap_node *node, size_t hash)
328 hmap_insert_fast (map, node, hash);
329 if (map->count > hmap_capacity (map))
330 hmap_reserve (map, map->count);
333 /* Inserts NODE into MAP with hash value HASH. Does not check
334 whether this causes MAP's current capacity to be exceeded.
335 The caller must take responsibility for that (or use
336 hmap_insert() instead).
338 This function runs in constant time.
340 This function does not verify that MAP does not already
341 contain a data item with the same value as NODE. If
342 duplicates should be disallowed (which is the usual case),
343 then the client must check for duplicates itself before
344 inserting the new node. */
346 hmap_insert_fast (struct hmap *map, struct hmap_node *node, size_t hash)
348 struct hmap_node **bucket = &map->buckets[hash & map->mask];
350 node->next = *bucket;
355 /* Removes NODE from MAP. The client is responsible for freeing
356 any data associated with NODE, if necessary.
358 Assuming uniform hashing, this function runs in constant time.
359 (Its runtime is proportional to the position of NODE in its
360 hash chain, so given a pathological hash function, e.g. one
361 that returns a constant value, its runtime degenerates to
362 linear in the length of NODE's hash chain.)
364 This function never reduces the number of buckets in MAP.
365 When one deletes a large number of nodes from a hash table,
366 calling hmap_shrink() afterward may therefore save a small
367 amount of memory. It is also more expensive to iterate
368 through a very sparse hash table than a denser one, so
369 shrinking the hash table could also save some time. However,
370 rehashing has an immediate cost that must be weighed against
373 hmap_delete() does not change NODE's hash value reported by
376 hmap_delete (struct hmap *map, struct hmap_node *node)
378 struct hmap_node **bucket = &map->buckets[node->hash & map->mask];
379 while (*bucket != node)
380 bucket = &(*bucket)->next;
381 *bucket = (*bucket)->next;
385 /* Returns the first node in MAP, or a null pointer if MAP is
388 Amortized over iterating through every data element in MAP,
389 this function runs in constant time. However, this assumes
390 that MAP is not excessively sparse, that is, that
391 hmap_capacity(MAP) is at most a constant factor greater than
392 hmap_count(MAP). This will always be true unless many nodes
393 have been inserted into MAP and then most or all of them
394 deleted; in such a case, calling hmap_shrink() is advised.
396 Nodes are returned in arbitrary order that may change whenever
397 the hash table's current capacity changes, as reported by
398 hmap_capacity(). Calls to hmap_insert(), hmap_reserve(), and
399 hmap_shrink() can change the capacity of a hash map.
400 Inserting a node with hmap_insert_fast() or deleting one with
401 hmap_delete() will not change the relative ordering of nodes.
403 The HMAP_FOR_EACH and HMAP_FOR_EACH_SAFE macros provide
404 convenient ways to iterate over all the nodes in a hash map.
405 The HMAP_FIRST macro is an interface to this particular
406 function that is often more convenient. */
407 static inline struct hmap_node *
408 hmap_first (const struct hmap *map)
410 return hmap_first_nonempty_bucket__ (map, 0);
413 /* Returns the next node in MAP following NODE, or a null pointer
414 if NODE is the last node in MAP.
416 Amortized over iterating through every data element in MAP,
417 this function runs in constant time. However, this assumes
418 that MAP is not excessively sparse, that is, that
419 hmap_capacity(MAP) is at most a constant factor greater than
420 hmap_count(MAP). This will always be true unless many nodes
421 have been inserted into MAP and then most or all of them
422 deleted; in such a case, calling hmap_shrink() is advised.
424 Nodes are returned in arbitrary order that may change whenever
425 the hash table's current capacity changes, as reported by
426 hmap_capacity(). Calls to hmap_insert(), hmap_reserve(), and
427 hmap_shrink() can change the capacity of a hash map.
428 Inserting a node with hmap_insert_fast() or deleting one with
429 hmap_delete() will not change the relative ordering of nodes.
431 The HMAP_FOR_EACH and HMAP_FOR_EACH_SAFE macros provide
432 convenient ways to iterate over all the nodes in a hash map.
433 The HMAP_NEXT macro is an interface to this particular
434 function that is often more convenient. */
435 static inline struct hmap_node *
436 hmap_next (const struct hmap *map, const struct hmap_node *node)
438 return (node->next != NULL
440 : hmap_first_nonempty_bucket__ (map, (node->hash & map->mask) + 1));
443 /* Returns the number of data items currently in MAP. */
445 hmap_count (const struct hmap *map)
450 /* Returns the current capacity of MAP, that is, the maximum
451 number of data elements that MAP may hold before it becomes
454 The capacity is advisory only: it is possible to insert any
455 number of data elements into a hash map regardless of its
456 capacity. However, inserting many more elements than the
457 map's capacity will degrade search performance. */
459 hmap_capacity (const struct hmap *map)
461 return hmap_mask_to_capacity__ (map->mask);
464 /* Implementation details. */
466 /* Returns the first node at or after NODE in NODE's chain that
467 has hash value HASH. */
468 static inline struct hmap_node *
469 hmap_find_hash__ (struct hmap_node *node, size_t hash)
471 for (; node != NULL; node = node->next)
472 if (node->hash == hash)
477 /* Returns the first node in the lowest-numbered nonempty bucket
478 in MAP whose index is START or higher, or a null pointer if
479 all such buckets are empty. */
480 static inline struct hmap_node *
481 hmap_first_nonempty_bucket__ (const struct hmap *map, size_t start)
485 for (i = start; i <= map->mask; i++)
486 if (map->buckets[i] != NULL)
487 return map->buckets[i];
491 /* Returns the hash table capacity associated with a given MASK,
492 which should be a value for the "mask" member of struct hmap.
493 MASK must be a power of 2 minus 1 (including 0), that is, its
494 value in binary must be all 1-bits. */
496 hmap_mask_to_capacity__ (size_t mask)
498 return (mask + 1) * 2;
501 /* Helper for HMAP_NULLABLE_DATA (to avoid evaluating its NODE
502 argument more than once). */
504 hmap_nullable_data__ (struct hmap_node *node, size_t member_offset)
506 return node != NULL ? (char *) node - member_offset : NULL;
509 #endif /* libpspp/hmap.h */