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