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     int j, added = 0;
 222
 223     set = lookupKeyWrite(c->db,c->argv[1]);
 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
 234     for (j = 2; j < c->argc; j++) {
 235         c->argv[j] = tryObjectEncoding(c->argv[j]);
 236         if (setTypeAdd(set,c->argv[j])) added++;
 237     }
 238     if (added) signalModifiedKey(c->db,c->argv[1]);
 239     server.dirty += added;
 240     addReplyLongLong(c,added);
 241 }
 242
 243 void sremCommand(redisClient *c) {
 244     robj *set;
 245     int j, deleted = 0;
 246
 247     if ((set = lookupKeyWriteOrReply(c,c->argv[1],shared.czero)) == NULL ||
 248         checkType(c,set,REDIS_SET)) return;
 249
 250     for (j = 2; j < c->argc; j++) {
 251         if (setTypeRemove(set,c->argv[j])) {
 252             if (setTypeSize(set) == 0) dbDelete(c->db,c->argv[1]);
 253             deleted++;
 254         }
 255     }
 256     if (deleted) {
 257         signalModifiedKey(c->db,c->argv[1]);
 258         server.dirty += deleted;
 259     }
 260     addReplyLongLong(c,deleted);
 261 }
 262
 263 void smoveCommand(redisClient *c) {
 264     robj *srcset, *dstset, *ele;
 265     srcset = lookupKeyWrite(c->db,c->argv[1]);
 266     dstset = lookupKeyWrite(c->db,c->argv[2]);
 267     ele = c->argv[3] = tryObjectEncoding(c->argv[3]);
 268
 269     /* If the source key does not exist return 0 */
 270     if (srcset == NULL) {
 271         addReply(c,shared.czero);
 272         return;
 273     }
 274
 275     /* If the source key has the wrong type, or the destination key
 276      * is set and has the wrong type, return with an error. */
 277     if (checkType(c,srcset,REDIS_SET) ||
 278         (dstset && checkType(c,dstset,REDIS_SET))) return;
 279
 280     /* If srcset and dstset are equal, SMOVE is a no-op */
 281     if (srcset == dstset) {
 282         addReply(c,shared.cone);
 283         return;
 284     }
 285
 286     /* If the element cannot be removed from the src set, return 0. */
 287     if (!setTypeRemove(srcset,ele)) {
 288         addReply(c,shared.czero);
 289         return;
 290     }
 291
 292     /* Remove the src set from the database when empty */
 293     if (setTypeSize(srcset) == 0) dbDelete(c->db,c->argv[1]);
 294     signalModifiedKey(c->db,c->argv[1]);
 295     signalModifiedKey(c->db,c->argv[2]);
 296     server.dirty++;
 297
 298     /* Create the destination set when it doesn't exist */
 299     if (!dstset) {
 300         dstset = setTypeCreate(ele);
 301         dbAdd(c->db,c->argv[2],dstset);
 302     }
 303
 304     /* An extra key has changed when ele was successfully added to dstset */
 305     if (setTypeAdd(dstset,ele)) server.dirty++;
 306     addReply(c,shared.cone);
 307 }
 308
 309 void sismemberCommand(redisClient *c) {
 310     robj *set;
 311
 312     if ((set = lookupKeyReadOrReply(c,c->argv[1],shared.czero)) == NULL ||
 313         checkType(c,set,REDIS_SET)) return;
 314
 315     c->argv[2] = tryObjectEncoding(c->argv[2]);
 316     if (setTypeIsMember(set,c->argv[2]))
 317         addReply(c,shared.cone);
 318     else
 319         addReply(c,shared.czero);
 320 }
 321
 322 void scardCommand(redisClient *c) {
 323     robj *o;
 324
 325     if ((o = lookupKeyReadOrReply(c,c->argv[1],shared.czero)) == NULL ||
 326         checkType(c,o,REDIS_SET)) return;
 327
 328     addReplyLongLong(c,setTypeSize(o));
 329 }
 330
 331 void spopCommand(redisClient *c) {
 332     robj *set, *ele;
 333     int64_t llele;
 334     int encoding;
 335
 336     if ((set = lookupKeyWriteOrReply(c,c->argv[1],shared.nullbulk)) == NULL ||
 337         checkType(c,set,REDIS_SET)) return;
 338
 339     encoding = setTypeRandomElement(set,&ele,&llele);
 340     if (encoding == REDIS_ENCODING_INTSET) {
 341         ele = createStringObjectFromLongLong(llele);
 342         set->ptr = intsetRemove(set->ptr,llele,NULL);
 343     } else {
 344         incrRefCount(ele);
 345         setTypeRemove(set,ele);
 346     }
 347
 348     /* Change argv to replicate as SREM */
 349     c->argc = 3;
 350     c->argv = zrealloc(c->argv,sizeof(robj*)*(c->argc));
 351
 352     /* Overwrite SREM with SPOP (same length) */
 353     redisAssert(sdslen(c->argv[0]->ptr) == 4);
 354     memcpy(c->argv[0]->ptr, "SREM", 4);
 355
 356     /* Popped element already has incremented refcount */
 357     c->argv[2] = ele;
 358
 359     addReplyBulk(c,ele);
 360     if (setTypeSize(set) == 0) dbDelete(c->db,c->argv[1]);
 361     signalModifiedKey(c->db,c->argv[1]);
 362     server.dirty++;
 363 }
 364
 365 void srandmemberCommand(redisClient *c) {
 366     robj *set, *ele;
 367     int64_t llele;
 368     int encoding;
 369
 370     if ((set = lookupKeyReadOrReply(c,c->argv[1],shared.nullbulk)) == NULL ||
 371         checkType(c,set,REDIS_SET)) return;
 372
 373     encoding = setTypeRandomElement(set,&ele,&llele);
 374     if (encoding == REDIS_ENCODING_INTSET) {
 375         addReplyBulkLongLong(c,llele);
 376     } else {
 377         addReplyBulk(c,ele);
 378     }
 379 }
 380
 381 int qsortCompareSetsByCardinality(const void *s1, const void *s2) {
 382     return setTypeSize(*(robj**)s1)-setTypeSize(*(robj**)s2);
 383 }
 384
 385 void sinterGenericCommand(redisClient *c, robj **setkeys, unsigned long setnum, robj *dstkey) {
 386     robj **sets = zmalloc(sizeof(robj*)*setnum);
 387     setTypeIterator *si;
 388     robj *eleobj, *dstset = NULL;
 389     int64_t intobj;
 390     void *replylen = NULL;
 391     unsigned long j, cardinality = 0;
 392     int encoding;
 393
 394     for (j = 0; j < setnum; j++) {
 395         robj *setobj = dstkey ?
 396             lookupKeyWrite(c->db,setkeys[j]) :
 397             lookupKeyRead(c->db,setkeys[j]);
 398         if (!setobj) {
 399             zfree(sets);
 400             if (dstkey) {
 401                 if (dbDelete(c->db,dstkey)) {
 402                     signalModifiedKey(c->db,dstkey);
 403                     server.dirty++;
 404                 }
 405                 addReply(c,shared.czero);
 406             } else {
 407                 addReply(c,shared.emptymultibulk);
 408             }
 409             return;
 410         }
 411         if (checkType(c,setobj,REDIS_SET)) {
 412             zfree(sets);
 413             return;
 414         }
 415         sets[j] = setobj;
 416     }
 417     /* Sort sets from the smallest to largest, this will improve our
 418      * algorithm's performace */
 419     qsort(sets,setnum,sizeof(robj*),qsortCompareSetsByCardinality);
 420
 421     /* The first thing we should output is the total number of elements...
 422      * since this is a multi-bulk write, but at this stage we don't know
 423      * the intersection set size, so we use a trick, append an empty object
 424      * to the output list and save the pointer to later modify it with the
 425      * right length */
 426     if (!dstkey) {
 427         replylen = addDeferredMultiBulkLength(c);
 428     } else {
 429         /* If we have a target key where to store the resulting set
 430          * create this key with an empty set inside */
 431         dstset = createIntsetObject();
 432     }
 433
 434     /* Iterate all the elements of the first (smallest) set, and test
 435      * the element against all the other sets, if at least one set does
 436      * not include the element it is discarded */
 437     si = setTypeInitIterator(sets[0]);
 438     while((encoding = setTypeNext(si,&eleobj,&intobj)) != -1) {
 439         for (j = 1; j < setnum; j++) {
 440             if (sets[j] == sets[0]) continue;
 441             if (encoding == REDIS_ENCODING_INTSET) {
 442                 /* intset with intset is simple... and fast */
 443                 if (sets[j]->encoding == REDIS_ENCODING_INTSET &&
 444                     !intsetFind((intset*)sets[j]->ptr,intobj))
 445                 {
 446                     break;
 447                 /* in order to compare an integer with an object we
 448                  * have to use the generic function, creating an object
 449                  * for this */
 450                 } else if (sets[j]->encoding == REDIS_ENCODING_HT) {
 451                     eleobj = createStringObjectFromLongLong(intobj);
 452                     if (!setTypeIsMember(sets[j],eleobj)) {
 453                         decrRefCount(eleobj);
 454                         break;
 455                     }
 456                     decrRefCount(eleobj);
 457                 }
 458             } else if (encoding == REDIS_ENCODING_HT) {
 459                 /* Optimization... if the source object is integer
 460                  * encoded AND the target set is an intset, we can get
 461                  * a much faster path. */
 462                 if (eleobj->encoding == REDIS_ENCODING_INT &&
 463                     sets[j]->encoding == REDIS_ENCODING_INTSET &&
 464                     !intsetFind((intset*)sets[j]->ptr,(long)eleobj->ptr))
 465                 {
 466                     break;
 467                 /* else... object to object check is easy as we use the
 468                  * type agnostic API here. */
 469                 } else if (!setTypeIsMember(sets[j],eleobj)) {
 470                     break;
 471                 }
 472             }
 473         }
 474
 475         /* Only take action when all sets contain the member */
 476         if (j == setnum) {
 477             if (!dstkey) {
 478                 if (encoding == REDIS_ENCODING_HT)
 479                     addReplyBulk(c,eleobj);
 480                 else
 481                     addReplyBulkLongLong(c,intobj);
 482                 cardinality++;
 483             } else {
 484                 if (encoding == REDIS_ENCODING_INTSET) {
 485                     eleobj = createStringObjectFromLongLong(intobj);
 486                     setTypeAdd(dstset,eleobj);
 487                     decrRefCount(eleobj);
 488                 } else {
 489                     setTypeAdd(dstset,eleobj);
 490                 }
 491             }
 492         }
 493     }
 494     setTypeReleaseIterator(si);
 495
 496     if (dstkey) {
 497         /* Store the resulting set into the target, if the intersection
 498          * is not an empty set. */
 499         dbDelete(c->db,dstkey);
 500         if (setTypeSize(dstset) > 0) {
 501             dbAdd(c->db,dstkey,dstset);
 502             addReplyLongLong(c,setTypeSize(dstset));
 503         } else {
 504             decrRefCount(dstset);
 505             addReply(c,shared.czero);
 506         }
 507         signalModifiedKey(c->db,dstkey);
 508         server.dirty++;
 509     } else {
 510         setDeferredMultiBulkLength(c,replylen,cardinality);
 511     }
 512     zfree(sets);
 513 }
 514
 515 void sinterCommand(redisClient *c) {
 516     sinterGenericCommand(c,c->argv+1,c->argc-1,NULL);
 517 }
 518
 519 void sinterstoreCommand(redisClient *c) {
 520     sinterGenericCommand(c,c->argv+2,c->argc-2,c->argv[1]);
 521 }
 522
 523 #define REDIS_OP_UNION 0
 524 #define REDIS_OP_DIFF 1
 525 #define REDIS_OP_INTER 2
 526
 527 void sunionDiffGenericCommand(redisClient *c, robj **setkeys, int setnum, robj *dstkey, int op) {
 528     robj **sets = zmalloc(sizeof(robj*)*setnum);
 529     setTypeIterator *si;
 530     robj *ele, *dstset = NULL;
 531     int j, cardinality = 0;
 532
 533     for (j = 0; j < setnum; j++) {
 534         robj *setobj = dstkey ?
 535             lookupKeyWrite(c->db,setkeys[j]) :
 536             lookupKeyRead(c->db,setkeys[j]);
 537         if (!setobj) {
 538             sets[j] = NULL;
 539             continue;
 540         }
 541         if (checkType(c,setobj,REDIS_SET)) {
 542             zfree(sets);
 543             return;
 544         }
 545         sets[j] = setobj;
 546     }
 547
 548     /* We need a temp set object to store our union. If the dstkey
 549      * is not NULL (that is, we are inside an SUNIONSTORE operation) then
 550      * this set object will be the resulting object to set into the target key*/
 551     dstset = createIntsetObject();
 552
 553     /* Iterate all the elements of all the sets, add every element a single
 554      * time to the result set */
 555     for (j = 0; j < setnum; j++) {
 556         if (op == REDIS_OP_DIFF && j == 0 && !sets[j]) break; /* result set is empty */
 557         if (!sets[j]) continue; /* non existing keys are like empty sets */
 558
 559         si = setTypeInitIterator(sets[j]);
 560         while((ele = setTypeNextObject(si)) != NULL) {
 561             if (op == REDIS_OP_UNION || j == 0) {
 562                 if (setTypeAdd(dstset,ele)) {
 563                     cardinality++;
 564                 }
 565             } else if (op == REDIS_OP_DIFF) {
 566                 if (setTypeRemove(dstset,ele)) {
 567                     cardinality--;
 568                 }
 569             }
 570             decrRefCount(ele);
 571         }
 572         setTypeReleaseIterator(si);
 573
 574         /* Exit when result set is empty. */
 575         if (op == REDIS_OP_DIFF && cardinality == 0) break;
 576     }
 577
 578     /* Output the content of the resulting set, if not in STORE mode */
 579     if (!dstkey) {
 580         addReplyMultiBulkLen(c,cardinality);
 581         si = setTypeInitIterator(dstset);
 582         while((ele = setTypeNextObject(si)) != NULL) {
 583             addReplyBulk(c,ele);
 584             decrRefCount(ele);
 585         }
 586         setTypeReleaseIterator(si);
 587         decrRefCount(dstset);
 588     } else {
 589         /* If we have a target key where to store the resulting set
 590          * create this key with the result set inside */
 591         dbDelete(c->db,dstkey);
 592         if (setTypeSize(dstset) > 0) {
 593             dbAdd(c->db,dstkey,dstset);
 594             addReplyLongLong(c,setTypeSize(dstset));
 595         } else {
 596             decrRefCount(dstset);
 597             addReply(c,shared.czero);
 598         }
 599         signalModifiedKey(c->db,dstkey);
 600         server.dirty++;
 601     }
 602     zfree(sets);
 603 }
 604
 605 void sunionCommand(redisClient *c) {
 606     sunionDiffGenericCommand(c,c->argv+1,c->argc-1,NULL,REDIS_OP_UNION);
 607 }
 608
 609 void sunionstoreCommand(redisClient *c) {
 610     sunionDiffGenericCommand(c,c->argv+2,c->argc-2,c->argv[1],REDIS_OP_UNION);
 611 }
 612
 613 void sdiffCommand(redisClient *c) {
 614     sunionDiffGenericCommand(c,c->argv+1,c->argc-1,NULL,REDIS_OP_DIFF);
 615 }
 616
 617 void sdiffstoreCommand(redisClient *c) {
 618     sunionDiffGenericCommand(c,c->argv+2,c->argc-2,c->argv[1],REDIS_OP_DIFF);
 619 }