SQLite

Check-in [d5b0269e0d]
Login

Many hyperlinks are disabled.
Use anonymous login to enable hyperlinks.

Overview
Comment:Optimizations in the hash table module. (CVS 1896)
Downloads: Tarball | ZIP archive
Timelines: family | ancestors | descendants | both | trunk
Files: files | file ages | folders
SHA1: d5b0269e0dd7c310460a7ffc4120ed45db823ce9
User & Date: drh 2004-08-20 14:08:51.000
Context
2004-08-20
16:02
Add support for named wildcards in SQL statements. (CVS 1897) (check-in: d3be0b7c5a user: drh tags: trunk)
14:08
Optimizations in the hash table module. (CVS 1896) (check-in: d5b0269e0d user: drh tags: trunk)
2004-08-19
15:12
Enhance lemon so that a @X instead of just X in the code expands to the major token value rather than the minor token value. Use this to make the parser a few hundred bytes smaller. (CVS 1895) (check-in: 28215096e0 user: drh tags: trunk)
Changes
Unified Diff Ignore Whitespace Patch
Changes to src/hash.c.
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
**    May you find forgiveness for yourself and forgive others.
**    May you share freely, never taking more than you give.
**
*************************************************************************
** This is the implementation of generic hash-tables
** used in SQLite.
**
** $Id: hash.c,v 1.14 2004/06/30 22:43:22 drh Exp $
*/
#include "sqliteInt.h"
#include <assert.h>

/* Turn bulk memory into a hash table object by initializing the
** fields of the Hash structure.
**
** "pNew" is a pointer to the hash table that is to be initialized.
** keyClass is one of the constants SQLITE_HASH_INT, SQLITE_HASH_POINTER,
** SQLITE_HASH_BINARY, or SQLITE_HASH_STRING.  The value of keyClass 
** determines what kind of key the hash table will use.  "copyKey" is
** true if the hash table should make its own private copy of keys and
** false if it should just use the supplied pointer.  CopyKey only makes
** sense for SQLITE_HASH_STRING and SQLITE_HASH_BINARY and is ignored
** for other key classes.
*/
void sqlite3HashInit(Hash *pNew, int keyClass, int copyKey){
  assert( pNew!=0 );
  assert( keyClass>=SQLITE_HASH_INT && keyClass<=SQLITE_HASH_BINARY );
  pNew->keyClass = keyClass;



  pNew->copyKey = copyKey &&
                (keyClass==SQLITE_HASH_STRING || keyClass==SQLITE_HASH_BINARY);
  pNew->first = 0;
  pNew->count = 0;
  pNew->htsize = 0;
  pNew->ht = 0;
}

/* Remove all entries from a hash table.  Reclaim all memory.







|


















|

>
>
>
|
<







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
**    May you find forgiveness for yourself and forgive others.
**    May you share freely, never taking more than you give.
**
*************************************************************************
** This is the implementation of generic hash-tables
** used in SQLite.
**
** $Id: hash.c,v 1.15 2004/08/20 14:08:51 drh Exp $
*/
#include "sqliteInt.h"
#include <assert.h>

/* Turn bulk memory into a hash table object by initializing the
** fields of the Hash structure.
**
** "pNew" is a pointer to the hash table that is to be initialized.
** keyClass is one of the constants SQLITE_HASH_INT, SQLITE_HASH_POINTER,
** SQLITE_HASH_BINARY, or SQLITE_HASH_STRING.  The value of keyClass 
** determines what kind of key the hash table will use.  "copyKey" is
** true if the hash table should make its own private copy of keys and
** false if it should just use the supplied pointer.  CopyKey only makes
** sense for SQLITE_HASH_STRING and SQLITE_HASH_BINARY and is ignored
** for other key classes.
*/
void sqlite3HashInit(Hash *pNew, int keyClass, int copyKey){
  assert( pNew!=0 );
  assert( keyClass>=SQLITE_HASH_STRING && keyClass<=SQLITE_HASH_BINARY );
  pNew->keyClass = keyClass;
#if 0
  if( keyClass==SQLITE_HASH_POINTER || keyClass==SQLITE_HASH_INT ) copyKey = 0;
#endif
  pNew->copyKey = copyKey;

  pNew->first = 0;
  pNew->count = 0;
  pNew->htsize = 0;
  pNew->ht = 0;
}

/* Remove all entries from a hash table.  Reclaim all memory.
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
/*
** Hash and comparison functions when the mode is SQLITE_HASH_STRING
*/
static int strHash(const void *pKey, int nKey){
  return sqlite3HashNoCase((const char*)pKey, nKey); 
}
static int strCompare(const void *pKey1, int n1, const void *pKey2, int n2){
  if( n1!=n2 ) return n2-n1;
  return sqlite3StrNICmp((const char*)pKey1,(const char*)pKey2,n1);
}

/*
** Hash and comparison functions when the mode is SQLITE_HASH_BINARY
*/
static int binHash(const void *pKey, int nKey){
  int h = 0;
  const char *z = (const char *)pKey;
  while( nKey-- > 0 ){
    h = (h<<3) ^ h ^ *(z++);
  }
  return h & 0x7fffffff;
}
static int binCompare(const void *pKey1, int n1, const void *pKey2, int n2){
  if( n1!=n2 ) return n2-n1;
  return memcmp(pKey1,pKey2,n1);
}

/*
** Return a pointer to the appropriate hash function given the key class.
**
** The C syntax in this function definition may be unfamilar to some 
** programmers, so we provide the following additional explanation:
**
** The name of the function is "hashFunction".  The function takes a
** single parameter "keyClass".  The return value of hashFunction()
** is a pointer to another function.  Specifically, the return value
** of hashFunction() is a pointer to a function that takes two parameters
** with types "const void*" and "int" and returns an "int".
*/
static int (*hashFunction(int keyClass))(const void*,int){

  switch( keyClass ){
    /* case SQLITE_HASH_INT:     return &intHash; // NOT USED */
    /* case SQLITE_HASH_POINTER: return &ptrHash; // NOT USED */
    case SQLITE_HASH_STRING:  return &strHash;
    case SQLITE_HASH_BINARY:  return &binHash;;
    default: break;
  }
  return 0;








}

/*
** Return a pointer to the appropriate hash function given the key class.
**
** For help in interpreted the obscure C code in the function definition,
** see the header comment on the previous function.
*/
static int (*compareFunction(int keyClass))(const void*,int,const void*,int){

  switch( keyClass ){
    /* case SQLITE_HASH_INT:     return &intCompare; // NOT USED */
    /* case SQLITE_HASH_POINTER: return &ptrCompare; // NOT USED */
    case SQLITE_HASH_STRING:  return &strCompare;
    case SQLITE_HASH_BINARY:  return &binCompare;
    default: break;
  }
  return 0;

































}


/* Resize the hash table so that it cantains "new_size" buckets.
** "new_size" must be a power of 2.  The hash table might fail 
** to resize if sqliteMalloc() fails.
*/
static void rehash(Hash *pH, int new_size){
  struct _ht *new_ht;            /* The new hash table */
  HashElem *elem, *next_elem;    /* For looping over existing elements */
  HashElem *x;                   /* Element being copied to new hash table */
  int (*xHash)(const void*,int); /* The hash function */

  assert( (new_size & (new_size-1))==0 );
  new_ht = (struct _ht *)sqliteMalloc( new_size*sizeof(struct _ht) );
  if( new_ht==0 ) return;
  if( pH->ht ) sqliteFree(pH->ht);
  pH->ht = new_ht;
  pH->htsize = new_size;
  xHash = hashFunction(pH->keyClass);
  for(elem=pH->first, pH->first=0; elem; elem = next_elem){
    int h = (*xHash)(elem->pKey, elem->nKey) & (new_size-1);
    next_elem = elem->next;
    x = new_ht[h].chain;
    if( x ){
      elem->next = x;
      elem->prev = x->prev;
      if( x->prev ) x->prev->next = elem;
      else          pH->first = elem;
      x->prev = elem;
    }else{
      elem->next = pH->first;
      if( pH->first ) pH->first->prev = elem;
      elem->prev = 0;
      pH->first = elem;
    }
    new_ht[h].chain = elem;
    new_ht[h].count++;
  }
}

/* This function (for internal use only) locates an element in an
** hash table that matches the given key.  The hash for this key has
** already been computed and is passed as the 4th parameter.
*/
static HashElem *findElementGivenHash(
  const Hash *pH,     /* The pH to be searched */
  const void *pKey,   /* The key we are searching for */
  int nKey,
  int h               /* The hash for this key. */
){
  HashElem *elem;                /* Used to loop thru the element list */
  int count;                     /* Number of elements left to test */
  int (*xCompare)(const void*,int,const void*,int);  /* comparison function */

  if( pH->ht ){

    elem = pH->ht[h].chain;
    count = pH->ht[h].count;
    xCompare = compareFunction(pH->keyClass);
    while( count-- && elem ){
      if( (*xCompare)(elem->pKey,elem->nKey,pKey,nKey)==0 ){ 
        return elem;
      }
      elem = elem->next;
    }
  }
  return 0;
}

/* Remove a single entry from the hash table given a pointer to that
** element and a hash on the element's key.
*/
static void removeElementGivenHash(
  Hash *pH,         /* The pH containing "elem" */
  HashElem* elem,   /* The element to be removed from the pH */
  int h             /* Hash value for the element */
){

  if( elem->prev ){
    elem->prev->next = elem->next; 
  }else{
    pH->first = elem->next;
  }
  if( elem->next ){
    elem->next->prev = elem->prev;
  }

  if( pH->ht[h].chain==elem ){
    pH->ht[h].chain = elem->next;
  }
  pH->ht[h].count--;
  if( pH->ht[h].count<=0 ){
    pH->ht[h].chain = 0;
  }
  if( pH->copyKey && elem->pKey ){
    sqliteFree(elem->pKey);
  }
  sqliteFree( elem );
  pH->count--;
}







|















|
















>

|
|





>
>
>
>
>
>
>
>









>

|
|





>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>










<












<
<
<
<
<
<
<
<
<
<
<
<
<
<
|


















>
|
|



















>








>
|
|

|
|
|







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
/*
** Hash and comparison functions when the mode is SQLITE_HASH_STRING
*/
static int strHash(const void *pKey, int nKey){
  return sqlite3HashNoCase((const char*)pKey, nKey); 
}
static int strCompare(const void *pKey1, int n1, const void *pKey2, int n2){
  if( n1!=n2 ) return 1;
  return sqlite3StrNICmp((const char*)pKey1,(const char*)pKey2,n1);
}

/*
** Hash and comparison functions when the mode is SQLITE_HASH_BINARY
*/
static int binHash(const void *pKey, int nKey){
  int h = 0;
  const char *z = (const char *)pKey;
  while( nKey-- > 0 ){
    h = (h<<3) ^ h ^ *(z++);
  }
  return h & 0x7fffffff;
}
static int binCompare(const void *pKey1, int n1, const void *pKey2, int n2){
  if( n1!=n2 ) return 1;
  return memcmp(pKey1,pKey2,n1);
}

/*
** Return a pointer to the appropriate hash function given the key class.
**
** The C syntax in this function definition may be unfamilar to some 
** programmers, so we provide the following additional explanation:
**
** The name of the function is "hashFunction".  The function takes a
** single parameter "keyClass".  The return value of hashFunction()
** is a pointer to another function.  Specifically, the return value
** of hashFunction() is a pointer to a function that takes two parameters
** with types "const void*" and "int" and returns an "int".
*/
static int (*hashFunction(int keyClass))(const void*,int){
#if 0  /* HASH_INT and HASH_POINTER are never used */
  switch( keyClass ){
    case SQLITE_HASH_INT:     return &intHash;
    case SQLITE_HASH_POINTER: return &ptrHash;
    case SQLITE_HASH_STRING:  return &strHash;
    case SQLITE_HASH_BINARY:  return &binHash;;
    default: break;
  }
  return 0;
#else
  if( keyClass==SQLITE_HASH_STRING ){
    return &strHash;
  }else{
    assert( keyClass==SQLITE_HASH_BINARY );
    return &binHash;
  }
#endif
}

/*
** Return a pointer to the appropriate hash function given the key class.
**
** For help in interpreted the obscure C code in the function definition,
** see the header comment on the previous function.
*/
static int (*compareFunction(int keyClass))(const void*,int,const void*,int){
#if 0 /* HASH_INT and HASH_POINTER are never used */
  switch( keyClass ){
    case SQLITE_HASH_INT:     return &intCompare;
    case SQLITE_HASH_POINTER: return &ptrCompare;
    case SQLITE_HASH_STRING:  return &strCompare;
    case SQLITE_HASH_BINARY:  return &binCompare;
    default: break;
  }
  return 0;
#else
  if( keyClass==SQLITE_HASH_STRING ){
    return &strCompare;
  }else{
    assert( keyClass==SQLITE_HASH_BINARY );
    return &binCompare;
  }
#endif
}

/* Link an element into the hash table
*/
static void insertElement(
  Hash *pH,              /* The complete hash table */
  struct _ht *pEntry,    /* The entry into which pNew is inserted */
  HashElem *pNew         /* The element to be inserted */
){
  HashElem *pHead;       /* First element already in pEntry */
  pHead = pEntry->chain;
  if( pHead ){
    pNew->next = pHead;
    pNew->prev = pHead->prev;
    if( pHead->prev ){ pHead->prev->next = pNew; }
    else             { pH->first = pNew; }
    pHead->prev = pNew;
  }else{
    pNew->next = pH->first;
    if( pH->first ){ pH->first->prev = pNew; }
    pNew->prev = 0;
    pH->first = pNew;
  }
  pEntry->count++;
  pEntry->chain = pNew;
}


/* Resize the hash table so that it cantains "new_size" buckets.
** "new_size" must be a power of 2.  The hash table might fail 
** to resize if sqliteMalloc() fails.
*/
static void rehash(Hash *pH, int new_size){
  struct _ht *new_ht;            /* The new hash table */
  HashElem *elem, *next_elem;    /* For looping over existing elements */

  int (*xHash)(const void*,int); /* The hash function */

  assert( (new_size & (new_size-1))==0 );
  new_ht = (struct _ht *)sqliteMalloc( new_size*sizeof(struct _ht) );
  if( new_ht==0 ) return;
  if( pH->ht ) sqliteFree(pH->ht);
  pH->ht = new_ht;
  pH->htsize = new_size;
  xHash = hashFunction(pH->keyClass);
  for(elem=pH->first, pH->first=0; elem; elem = next_elem){
    int h = (*xHash)(elem->pKey, elem->nKey) & (new_size-1);
    next_elem = elem->next;














    insertElement(pH, &new_ht[h], elem);
  }
}

/* This function (for internal use only) locates an element in an
** hash table that matches the given key.  The hash for this key has
** already been computed and is passed as the 4th parameter.
*/
static HashElem *findElementGivenHash(
  const Hash *pH,     /* The pH to be searched */
  const void *pKey,   /* The key we are searching for */
  int nKey,
  int h               /* The hash for this key. */
){
  HashElem *elem;                /* Used to loop thru the element list */
  int count;                     /* Number of elements left to test */
  int (*xCompare)(const void*,int,const void*,int);  /* comparison function */

  if( pH->ht ){
    struct _ht *pEntry = &pH->ht[h];
    elem = pEntry->chain;
    count = pEntry->count;
    xCompare = compareFunction(pH->keyClass);
    while( count-- && elem ){
      if( (*xCompare)(elem->pKey,elem->nKey,pKey,nKey)==0 ){ 
        return elem;
      }
      elem = elem->next;
    }
  }
  return 0;
}

/* Remove a single entry from the hash table given a pointer to that
** element and a hash on the element's key.
*/
static void removeElementGivenHash(
  Hash *pH,         /* The pH containing "elem" */
  HashElem* elem,   /* The element to be removed from the pH */
  int h             /* Hash value for the element */
){
  struct _ht *pEntry;
  if( elem->prev ){
    elem->prev->next = elem->next; 
  }else{
    pH->first = elem->next;
  }
  if( elem->next ){
    elem->next->prev = elem->prev;
  }
  pEntry = &pH->ht[h];
  if( pEntry->chain==elem ){
    pEntry->chain = elem->next;
  }
  pEntry->count--;
  if( pEntry->count<=0 ){
    pEntry->chain = 0;
  }
  if( pH->copyKey && elem->pKey ){
    sqliteFree(elem->pKey);
  }
  sqliteFree( elem );
  pH->count--;
}
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
    }
    memcpy((void*)new_elem->pKey, pKey, nKey);
  }else{
    new_elem->pKey = (void*)pKey;
  }
  new_elem->nKey = nKey;
  pH->count++;
  if( pH->htsize==0 ) rehash(pH,8);

  if( pH->htsize==0 ){
    pH->count = 0;
    sqliteFree(new_elem);
    return data;

  }
  if( pH->count > pH->htsize ){
    rehash(pH,pH->htsize*2);
  }

  assert( (pH->htsize & (pH->htsize-1))==0 );
  h = hraw & (pH->htsize-1);
  elem = pH->ht[h].chain;
  if( elem ){
    new_elem->next = elem;
    new_elem->prev = elem->prev;
    if( elem->prev ){ elem->prev->next = new_elem; }
    else            { pH->first = new_elem; }
    elem->prev = new_elem;
  }else{
    new_elem->next = pH->first;
    new_elem->prev = 0;
    if( pH->first ){ pH->first->prev = new_elem; }
    pH->first = new_elem;
  }
  pH->ht[h].count++;
  pH->ht[h].chain = new_elem;
  new_elem->data = data;
  return 0;
}







|
>
|
|
|
|
>




>


<
<
<
<
<
<
<
<
<
<
<
<
<
|
<



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
    }
    memcpy((void*)new_elem->pKey, pKey, nKey);
  }else{
    new_elem->pKey = (void*)pKey;
  }
  new_elem->nKey = nKey;
  pH->count++;
  if( pH->htsize==0 ){
    rehash(pH,8);
    if( pH->htsize==0 ){
      pH->count = 0;
      sqliteFree(new_elem);
      return data;
    }
  }
  if( pH->count > pH->htsize ){
    rehash(pH,pH->htsize*2);
  }
  assert( pH->htsize>0 );
  assert( (pH->htsize & (pH->htsize-1))==0 );
  h = hraw & (pH->htsize-1);













  insertElement(pH, &pH->ht[h], new_elem);

  new_elem->data = data;
  return 0;
}
Changes to src/hash.h.
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
**    May you find forgiveness for yourself and forgive others.
**    May you share freely, never taking more than you give.
**
*************************************************************************
** This is the header file for the generic hash-table implemenation
** used in SQLite.
**
** $Id: hash.h,v 1.7 2004/05/08 08:23:25 danielk1977 Exp $
*/
#ifndef _SQLITE_HASH_H_
#define _SQLITE_HASH_H_

/* Forward declarations of structures. */
typedef struct Hash Hash;
typedef struct HashElem HashElem;







|







8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
**    May you find forgiveness for yourself and forgive others.
**    May you share freely, never taking more than you give.
**
*************************************************************************
** This is the header file for the generic hash-table implemenation
** used in SQLite.
**
** $Id: hash.h,v 1.8 2004/08/20 14:08:51 drh Exp $
*/
#ifndef _SQLITE_HASH_H_
#define _SQLITE_HASH_H_

/* Forward declarations of structures. */
typedef struct Hash Hash;
typedef struct HashElem HashElem;
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
**
**   SQLITE_HASH_BINARY      pKey points to binary data nKey bytes long. 
**                           memcmp() is used to compare keys.
**
** A copy of the key is made for SQLITE_HASH_STRING and SQLITE_HASH_BINARY
** if the copyKey parameter to HashInit is 1.  
*/
#define SQLITE_HASH_INT       1
/* #define SQLITE_HASH_POINTER   2 // NOT USED */
#define SQLITE_HASH_STRING    3
#define SQLITE_HASH_BINARY    4

/*
** Access routines.  To delete, insert a NULL pointer.
*/







|







66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
**
**   SQLITE_HASH_BINARY      pKey points to binary data nKey bytes long. 
**                           memcmp() is used to compare keys.
**
** A copy of the key is made for SQLITE_HASH_STRING and SQLITE_HASH_BINARY
** if the copyKey parameter to HashInit is 1.  
*/
/* #define SQLITE_HASH_INT       1 // NOT USED */
/* #define SQLITE_HASH_POINTER   2 // NOT USED */
#define SQLITE_HASH_STRING    3
#define SQLITE_HASH_BINARY    4

/*
** Access routines.  To delete, insert a NULL pointer.
*/
103
104
105
106
107
108
109
110
111
112

/*
** Number of entries in a hash table
*/
#define sqliteHashCount(H)  ((H)->count)

#endif /* _SQLITE_HASH_H_ */










<
<
<
103
104
105
106
107
108
109




/*
** Number of entries in a hash table
*/
#define sqliteHashCount(H)  ((H)->count)

#endif /* _SQLITE_HASH_H_ */