src/t_set.c

   1 #include "redis.h"
   2
   3 /*-----------------------------------------------------------------------------
   4  * Set Commands
   5  *----------------------------------------------------------------------------*/
   6
   7 /* Factory method to return a set that *can* hold "value". When the object has
   8  * an integer-encodable value, an intset will be returned. Otherwise a regular
   9  * hash table. */
  10 robj *setTypeCreate(robj *value) {
  11     if (isObjectRepresentableAsLongLong(value,NULL) == REDIS_OK)
  12         return createIntsetObject();
  13     return createSetObject();
  14 }
  15
  16 int setTypeAdd(robj *subject, robj *value) {
  17     long long llval;
  18     if (subject->encoding == REDIS_ENCODING_HT) {
  19         if (dictAdd(subject->ptr,value,NULL) == DICT_OK) {
  20             incrRefCount(value);
  21             return 1;
  22         }
  23     } else if (subject->encoding == REDIS_ENCODING_INTSET) {
  24         if (isObjectRepresentableAsLongLong(value,&llval) == REDIS_OK) {
  25             uint8_t success = 0;
  26             subject->ptr = intsetAdd(subject->ptr,llval,&success);
  27             if (success) {
  28                 /* Convert to regular set when the intset contains
  29                  * too many entries. */
  30                 if (intsetLen(subject->ptr) > server.set_max_intset_entries)
  31                     setTypeConvert(subject,REDIS_ENCODING_HT);
  32                 return 1;
  33             }
  34         } else {
  35             /* Failed to get integer from object, convert to regular set. */
  36             setTypeConvert(subject,REDIS_ENCODING_HT);
  37
  38             /* The set *was* an intset and this value is not integer
  39              * encodable, so dictAdd should always work. */
  40             redisAssert(dictAdd(subject->ptr,value,NULL) == DICT_OK);
  41             incrRefCount(value);
  42             return 1;
  43         }
  44     } else {
  45         redisPanic("Unknown set encoding");
  46     }
  47     return 0;
  48 }
  49
  50 int setTypeRemove(robj *setobj, robj *value) {
  51     long long llval;
  52     if (setobj->encoding == REDIS_ENCODING_HT) {
  53         if (dictDelete(setobj->ptr,value) == DICT_OK) {
  54             if (htNeedsResize(setobj->ptr)) dictResize(setobj->ptr);
  55             return 1;
  56         }
  57     } else if (setobj->encoding == REDIS_ENCODING_INTSET) {
  58         if (isObjectRepresentableAsLongLong(value,&llval) == REDIS_OK) {
  59             int success;
  60             setobj->ptr = intsetRemove(setobj->ptr,llval,&success);
  61             if (success) return 1;
  62         }
  63     } else {
  64         redisPanic("Unknown set encoding");
  65     }
  66     return 0;
  67 }
  68
  69 int setTypeIsMember(robj *subject, robj *value) {
  70     long long llval;
  71     if (subject->encoding == REDIS_ENCODING_HT) {
  72         return dictFind((dict*)subject->ptr,value) != NULL;
  73     } else if (subject->encoding == REDIS_ENCODING_INTSET) {
  74         if (isObjectRepresentableAsLongLong(value,&llval) == REDIS_OK) {
  75             return intsetFind((intset*)subject->ptr,llval);
  76         }
  77     } else {
  78         redisPanic("Unknown set encoding");
  79     }
  80     return 0;
  81 }
  82
  83 setTypeIterator *setTypeInitIterator(robj *subject) {
  84     setTypeIterator *si = zmalloc(sizeof(setTypeIterator));
  85     si->subject = subject;
  86     si->encoding = subject->encoding;
  87     if (si->encoding == REDIS_ENCODING_HT) {
  88         si->di = dictGetIterator(subject->ptr);
  89     } else if (si->encoding == REDIS_ENCODING_INTSET) {
  90         si->ii = 0;
  91     } else {
  92         redisPanic("Unknown set encoding");
  93     }
  94     return si;
  95 }
  96
  97 void setTypeReleaseIterator(setTypeIterator *si) {
  98     if (si->encoding == REDIS_ENCODING_HT)
  99         dictReleaseIterator(si->di);
 100     zfree(si);
 101 }
 102
 103 /* Move to the next entry in the set. Returns the object at the current
 104  * position.
 105  *
 106  * Since set elements can be internally be stored as redis objects or
 107  * simple arrays of integers, setTypeNext returns the encoding of the
 108  * set object you are iterating, and will populate the appropriate pointer
 109  * (eobj) or (llobj) accordingly.
 110  *
 111  * When there are no longer elements -1 is returned.
 112  * Returned objects ref count is not incremented, so this function is
 113  * copy on write friendly. */
 114 int setTypeNext(setTypeIterator *si, robj **objele, int64_t *llele) {
 115     if (si->encoding == REDIS_ENCODING_HT) {
 116         dictEntry *de = dictNext(si->di);
 117         if (de == NULL) return -1;
 118         *objele = dictGetEntryKey(de);
 119     } else if (si->encoding == REDIS_ENCODING_INTSET) {
 120         if (!intsetGet(si->subject->ptr,si->ii++,llele))
 121             return -1;
 122     }
 123     return si->encoding;
 124 }
 125
 126 /* The not copy on write friendly version but easy to use version
 127  * of setTypeNext() is setTypeNextObject(), returning new objects
 128  * or incrementing the ref count of returned objects. So if you don't
 129  * retain a pointer to this object you should call decrRefCount() against it.
 130  *
 131  * This function is the way to go for write operations where COW is not
 132  * an issue as the result will be anyway of incrementing the ref count. */
 133 robj *setTypeNextObject(setTypeIterator *si) {
 134     int64_t intele;
 135     robj *objele;
 136     int encoding;
 137
 138     encoding = setTypeNext(si,&objele,&intele);
 139     switch(encoding) {
 140         case -1:    return NULL;
 141         case REDIS_ENCODING_INTSET:
 142             return createStringObjectFromLongLong(intele);
 143         case REDIS_ENCODING_HT:
 144             incrRefCount(objele);
 145             return objele;
 146         default:
 147             redisPanic("Unsupported encoding");
 148     }
 149     return NULL; /* just to suppress warnings */
 150 }
 151
 152 /* Return random element from a non empty set.
 153  * The returned element can be a int64_t value if the set is encoded
 154  * as an "intset" blob of integers, or a redis object if the set
 155  * is a regular set.
 156  *
 157  * The caller provides both pointers to be populated with the right
 158  * object. The return value of the function is the object->encoding
 159  * field of the object and is used by the caller to check if the
 160  * int64_t pointer or the redis object pointere was populated.
 161  *
 162  * When an object is returned (the set was a real set) the ref count
 163  * of the object is not incremented so this function can be considered
 164  * copy on write friendly. */
 165 int setTypeRandomElement(robj *setobj, robj **objele, int64_t *llele) {
 166     if (setobj->encoding == REDIS_ENCODING_HT) {
 167         dictEntry *de = dictGetRandomKey(setobj->ptr);
 168         *objele = dictGetEntryKey(de);
 169     } else if (setobj->encoding == REDIS_ENCODING_INTSET) {
 170         *llele = intsetRandom(setobj->ptr);
 171     } else {
 172         redisPanic("Unknown set encoding");
 173     }
 174     return setobj->encoding;
 175 }
 176
 177 unsigned long setTypeSize(robj *subject) {
 178     if (subject->encoding == REDIS_ENCODING_HT) {
 179         return dictSize((dict*)subject->ptr);
 180     } else if (subject->encoding == REDIS_ENCODING_INTSET) {
 181         return intsetLen((intset*)subject->ptr);
 182     } else {
 183         redisPanic("Unknown set encoding");
 184     }
 185 }
 186
 187 /* Convert the set to specified encoding. The resulting dict (when converting
 188  * to a hashtable) is presized to hold the number of elements in the original
 189  * set. */
 190 void setTypeConvert(robj *setobj, int enc) {
 191     setTypeIterator *si;
 192     redisAssert(setobj->type == REDIS_SET &&
 193                 setobj->encoding == REDIS_ENCODING_INTSET);
 194
 195     if (enc == REDIS_ENCODING_HT) {
 196         int64_t intele;
 197         dict *d = dictCreate(&setDictType,NULL);
 198         robj *element;
 199
 200         /* Presize the dict to avoid rehashing */
 201         dictExpand(d,intsetLen(setobj->ptr));
 202
 203         /* To add the elements we extract integers and create redis objects */
 204         si = setTypeInitIterator(setobj);
 205         while (setTypeNext(si,NULL,&intele) != -1) {
 206             element = createStringObjectFromLongLong(intele);
 207             redisAssert(dictAdd(d,element,NULL) == DICT_OK);
 208         }
 209         setTypeReleaseIterator(si);
 210
 211         setobj->encoding = REDIS_ENCODING_HT;
 212         zfree(setobj->ptr);
 213         setobj->ptr = d;
 214     } else {
 215         redisPanic("Unsupported set conversion");
 216     }
 217 }
 218
 219 void saddCommand(redisClient *c) {
 220     robj *set;
 221
 222     set = lookupKeyWrite(c->db,c->argv[1]);
 223     c->argv[2] = tryObjectEncoding(c->argv[2]);
 224     if (set == NULL) {
 225         set = setTypeCreate(c->argv[2]);
 226         dbAdd(c->db,c->argv[1],set);
 227     } else {
 228         if (set->type != REDIS_SET) {
 229             addReply(c,shared.wrongtypeerr);
 230             return;
 231         }
 232     }
 233     if (setTypeAdd(set,c->argv[2])) {
 234         signalModifiedKey(c->db,c->argv[1]);
 235         server.dirty++;
 236         addReply(c,shared.cone);
 237     } else {
 238         addReply(c,shared.czero);
 239     }
 240 }
 241
 242 void sremCommand(redisClient *c) {
 243     robj *set;
 244
 245     if ((set = lookupKeyWriteOrReply(c,c->argv[1],shared.czero)) == NULL ||
 246         checkType(c,set,REDIS_SET)) return;
 247
 248     c->argv[2] = tryObjectEncoding(c->argv[2]);
 249     if (setTypeRemove(set,c->argv[2])) {
 250         if (setTypeSize(set) == 0) dbDelete(c->db,c->argv[1]);
 251         signalModifiedKey(c->db,c->argv[1]);
 252         server.dirty++;
 253         addReply(c,shared.cone);
 254     } else {
 255         addReply(c,shared.czero);
 256     }
 257 }
 258
 259 void smoveCommand(redisClient *c) {
 260     robj *srcset, *dstset, *ele;
 261     srcset = lookupKeyWrite(c->db,c->argv[1]);
 262     dstset = lookupKeyWrite(c->db,c->argv[2]);
 263     ele = c->argv[3] = tryObjectEncoding(c->argv[3]);
 264
 265     /* If the source key does not exist return 0 */
 266     if (srcset == NULL) {
 267         addReply(c,shared.czero);
 268         return;
 269     }
 270
 271     /* If the source key has the wrong type, or the destination key
 272      * is set and has the wrong type, return with an error. */
 273     if (checkType(c,srcset,REDIS_SET) ||
 274         (dstset && checkType(c,dstset,REDIS_SET))) return;
 275
 276     /* If srcset and dstset are equal, SMOVE is a no-op */
 277     if (srcset == dstset) {
 278         addReply(c,shared.cone);
 279         return;
 280     }
 281
 282     /* If the element cannot be removed from the src set, return 0. */
 283     if (!setTypeRemove(srcset,ele)) {
 284         addReply(c,shared.czero);
 285         return;
 286     }
 287
 288     /* Remove the src set from the database when empty */
 289     if (setTypeSize(srcset) == 0) dbDelete(c->db,c->argv[1]);
 290     signalModifiedKey(c->db,c->argv[1]);
 291     signalModifiedKey(c->db,c->argv[2]);
 292     server.dirty++;
 293
 294     /* Create the destination set when it doesn't exist */
 295     if (!dstset) {
 296         dstset = setTypeCreate(ele);
 297         dbAdd(c->db,c->argv[2],dstset);
 298     }
 299
 300     /* An extra key has changed when ele was successfully added to dstset */
 301     if (setTypeAdd(dstset,ele)) server.dirty++;
 302     addReply(c,shared.cone);
 303 }
 304
 305 void sismemberCommand(redisClient *c) {
 306     robj *set;
 307
 308     if ((set = lookupKeyReadOrReply(c,c->argv[1],shared.czero)) == NULL ||
 309         checkType(c,set,REDIS_SET)) return;
 310
 311     c->argv[2] = tryObjectEncoding(c->argv[2]);
 312     if (setTypeIsMember(set,c->argv[2]))
 313         addReply(c,shared.cone);
 314     else
 315         addReply(c,shared.czero);
 316 }
 317
 318 void scardCommand(redisClient *c) {
 319     robj *o;
 320
 321     if ((o = lookupKeyReadOrReply(c,c->argv[1],shared.czero)) == NULL ||
 322         checkType(c,o,REDIS_SET)) return;
 323
 324     addReplyLongLong(c,setTypeSize(o));
 325 }
 326
 327 void spopCommand(redisClient *c) {
 328     robj *set, *ele;
 329     int64_t llele;
 330     int encoding;
 331
 332     if ((set = lookupKeyWriteOrReply(c,c->argv[1],shared.nullbulk)) == NULL ||
 333         checkType(c,set,REDIS_SET)) return;
 334
 335     encoding = setTypeRandomElement(set,&ele,&llele);
 336     if (encoding == REDIS_ENCODING_INTSET) {
 337         addReplyBulkLongLong(c,llele);
 338         set->ptr = intsetRemove(set->ptr,llele,NULL);
 339     } else {
 340         addReplyBulk(c,ele);
 341         setTypeRemove(set,ele);
 342     }
 343     if (setTypeSize(set) == 0) dbDelete(c->db,c->argv[1]);
 344     signalModifiedKey(c->db,c->argv[1]);
 345     server.dirty++;
 346 }
 347
 348 void srandmemberCommand(redisClient *c) {
 349     robj *set, *ele;
 350     int64_t llele;
 351     int encoding;
 352
 353     if ((set = lookupKeyReadOrReply(c,c->argv[1],shared.nullbulk)) == NULL ||
 354         checkType(c,set,REDIS_SET)) return;
 355
 356     encoding = setTypeRandomElement(set,&ele,&llele);
 357     if (encoding == REDIS_ENCODING_INTSET) {
 358         addReplyBulkLongLong(c,llele);
 359     } else {
 360         addReplyBulk(c,ele);
 361     }
 362 }
 363
 364 int qsortCompareSetsByCardinality(const void *s1, const void *s2) {
 365     return setTypeSize(*(robj**)s1)-setTypeSize(*(robj**)s2);
 366 }
 367
 368 void sinterGenericCommand(redisClient *c, robj **setkeys, unsigned long setnum, robj *dstkey) {
 369     robj **sets = zmalloc(sizeof(robj*)*setnum);
 370     setTypeIterator *si;
 371     robj *eleobj, *dstset = NULL;
 372     int64_t intobj;
 373     void *replylen = NULL;
 374     unsigned long j, cardinality = 0;
 375     int encoding;
 376
 377     for (j = 0; j < setnum; j++) {
 378         robj *setobj = dstkey ?
 379             lookupKeyWrite(c->db,setkeys[j]) :
 380             lookupKeyRead(c->db,setkeys[j]);
 381         if (!setobj) {
 382             zfree(sets);
 383             if (dstkey) {
 384                 if (dbDelete(c->db,dstkey)) {
 385                     signalModifiedKey(c->db,dstkey);
 386                     server.dirty++;
 387                 }
 388                 addReply(c,shared.czero);
 389             } else {
 390                 addReply(c,shared.emptymultibulk);
 391             }
 392             return;
 393         }
 394         if (checkType(c,setobj,REDIS_SET)) {
 395             zfree(sets);
 396             return;
 397         }
 398         sets[j] = setobj;
 399     }
 400     /* Sort sets from the smallest to largest, this will improve our
 401      * algorithm's performace */
 402     qsort(sets,setnum,sizeof(robj*),qsortCompareSetsByCardinality);
 403
 404     /* The first thing we should output is the total number of elements...
 405      * since this is a multi-bulk write, but at this stage we don't know
 406      * the intersection set size, so we use a trick, append an empty object
 407      * to the output list and save the pointer to later modify it with the
 408      * right length */
 409     if (!dstkey) {
 410         replylen = addDeferredMultiBulkLength(c);
 411     } else {
 412         /* If we have a target key where to store the resulting set
 413          * create this key with an empty set inside */
 414         dstset = createIntsetObject();
 415     }
 416
 417     /* Iterate all the elements of the first (smallest) set, and test
 418      * the element against all the other sets, if at least one set does
 419      * not include the element it is discarded */
 420     si = setTypeInitIterator(sets[0]);
 421     while((encoding = setTypeNext(si,&eleobj,&intobj)) != -1) {
 422         for (j = 1; j < setnum; j++) {
 423             if (encoding == REDIS_ENCODING_INTSET) {
 424                 /* intset with intset is simple... and fast */
 425                 if (sets[j]->encoding == REDIS_ENCODING_INTSET &&
 426                     !intsetFind((intset*)sets[j]->ptr,intobj))
 427                 {
 428                     break;
 429                 /* in order to compare an integer with an object we
 430                  * have to use the generic function, creating an object
 431                  * for this */
 432                 } else if (sets[j]->encoding == REDIS_ENCODING_HT) {
 433                     eleobj = createStringObjectFromLongLong(intobj);
 434                     if (!setTypeIsMember(sets[j],eleobj)) {
 435                         decrRefCount(eleobj);
 436                         break;
 437                     }
 438                     decrRefCount(eleobj);
 439                 }
 440             } else if (encoding == REDIS_ENCODING_HT) {
 441                 /* Optimization... if the source object is integer
 442                  * encoded AND the target set is an intset, we can get
 443                  * a much faster path. */
 444                 if (eleobj->encoding == REDIS_ENCODING_INT &&
 445                     sets[j]->encoding == REDIS_ENCODING_INTSET &&
 446                     !intsetFind((intset*)sets[j]->ptr,(long)eleobj->ptr))
 447                 {
 448                     break;
 449                 /* else... object to object check is easy as we use the
 450                  * type agnostic API here. */
 451                 } else if (!setTypeIsMember(sets[j],eleobj)) {
 452                     break;
 453                 }
 454             }
 455         }
 456
 457         /* Only take action when all sets contain the member */
 458         if (j == setnum) {
 459             if (!dstkey) {
 460                 if (encoding == REDIS_ENCODING_HT)
 461                     addReplyBulk(c,eleobj);
 462                 else
 463                     addReplyBulkLongLong(c,intobj);
 464                 cardinality++;
 465             } else {
 466                 if (encoding == REDIS_ENCODING_INTSET) {
 467                     eleobj = createStringObjectFromLongLong(intobj);
 468                     setTypeAdd(dstset,eleobj);
 469                     decrRefCount(eleobj);
 470                 } else {
 471                     setTypeAdd(dstset,eleobj);
 472                 }
 473             }
 474         }
 475     }
 476     setTypeReleaseIterator(si);
 477
 478     if (dstkey) {
 479         /* Store the resulting set into the target, if the intersection
 480          * is not an empty set. */
 481         dbDelete(c->db,dstkey);
 482         if (setTypeSize(dstset) > 0) {
 483             dbAdd(c->db,dstkey,dstset);
 484             addReplyLongLong(c,setTypeSize(dstset));
 485         } else {
 486             decrRefCount(dstset);
 487             addReply(c,shared.czero);
 488         }
 489         signalModifiedKey(c->db,dstkey);
 490         server.dirty++;
 491     } else {
 492         setDeferredMultiBulkLength(c,replylen,cardinality);
 493     }
 494     zfree(sets);
 495 }
 496
 497 void sinterCommand(redisClient *c) {
 498     sinterGenericCommand(c,c->argv+1,c->argc-1,NULL);
 499 }
 500
 501 void sinterstoreCommand(redisClient *c) {
 502     sinterGenericCommand(c,c->argv+2,c->argc-2,c->argv[1]);
 503 }
 504
 505 #define REDIS_OP_UNION 0
 506 #define REDIS_OP_DIFF 1
 507 #define REDIS_OP_INTER 2
 508
 509 void sunionDiffGenericCommand(redisClient *c, robj **setkeys, int setnum, robj *dstkey, int op) {
 510     robj **sets = zmalloc(sizeof(robj*)*setnum);
 511     setTypeIterator *si;
 512     robj *ele, *dstset = NULL;
 513     int j, cardinality = 0;
 514
 515     for (j = 0; j < setnum; j++) {
 516         robj *setobj = dstkey ?
 517             lookupKeyWrite(c->db,setkeys[j]) :
 518             lookupKeyRead(c->db,setkeys[j]);
 519         if (!setobj) {
 520             sets[j] = NULL;
 521             continue;
 522         }
 523         if (checkType(c,setobj,REDIS_SET)) {
 524             zfree(sets);
 525             return;
 526         }
 527         sets[j] = setobj;
 528     }
 529
 530     /* We need a temp set object to store our union. If the dstkey
 531      * is not NULL (that is, we are inside an SUNIONSTORE operation) then
 532      * this set object will be the resulting object to set into the target key*/
 533     dstset = createIntsetObject();
 534
 535     /* Iterate all the elements of all the sets, add every element a single
 536      * time to the result set */
 537     for (j = 0; j < setnum; j++) {
 538         if (op == REDIS_OP_DIFF && j == 0 && !sets[j]) break; /* result set is empty */
 539         if (!sets[j]) continue; /* non existing keys are like empty sets */
 540
 541         si = setTypeInitIterator(sets[j]);
 542         while((ele = setTypeNextObject(si)) != NULL) {
 543             if (op == REDIS_OP_UNION || j == 0) {
 544                 if (setTypeAdd(dstset,ele)) {
 545                     cardinality++;
 546                 }
 547             } else if (op == REDIS_OP_DIFF) {
 548                 if (setTypeRemove(dstset,ele)) {
 549                     cardinality--;
 550                 }
 551             }
 552             decrRefCount(ele);
 553         }
 554         setTypeReleaseIterator(si);
 555
 556         /* Exit when result set is empty. */
 557         if (op == REDIS_OP_DIFF && cardinality == 0) break;
 558     }
 559
 560     /* Output the content of the resulting set, if not in STORE mode */
 561     if (!dstkey) {
 562         addReplyMultiBulkLen(c,cardinality);
 563         si = setTypeInitIterator(dstset);
 564         while((ele = setTypeNextObject(si)) != NULL) {
 565             addReplyBulk(c,ele);
 566             decrRefCount(ele);
 567         }
 568         setTypeReleaseIterator(si);
 569         decrRefCount(dstset);
 570     } else {
 571         /* If we have a target key where to store the resulting set
 572          * create this key with the result set inside */
 573         dbDelete(c->db,dstkey);
 574         if (setTypeSize(dstset) > 0) {
 575             dbAdd(c->db,dstkey,dstset);
 576             addReplyLongLong(c,setTypeSize(dstset));
 577         } else {
 578             decrRefCount(dstset);
 579             addReply(c,shared.czero);
 580         }
 581         signalModifiedKey(c->db,dstkey);
 582         server.dirty++;
 583     }
 584     zfree(sets);
 585 }
 586
 587 void sunionCommand(redisClient *c) {
 588     sunionDiffGenericCommand(c,c->argv+1,c->argc-1,NULL,REDIS_OP_UNION);
 589 }
 590
 591 void sunionstoreCommand(redisClient *c) {
 592     sunionDiffGenericCommand(c,c->argv+2,c->argc-2,c->argv[1],REDIS_OP_UNION);
 593 }
 594
 595 void sdiffCommand(redisClient *c) {
 596     sunionDiffGenericCommand(c,c->argv+1,c->argc-1,NULL,REDIS_OP_DIFF);
 597 }
 598
 599 void sdiffstoreCommand(redisClient *c) {
 600     sunionDiffGenericCommand(c,c->argv+2,c->argc-2,c->argv[1],REDIS_OP_DIFF);
 601 }