fts1.c

sqlite-3.4.1,嵌入式数据库.是一个功能强大的开源数据库,给学习和研发以及小型公司的发展带来了全所未有的好处.

  return pReader->iLastPos;
}

/* Skip past the end of a position list. */
static void skipPositionList(DocListReader *pReader){
  DocList *p = pReader->pDoclist;
  if( p && p->iType>=DL_POSITIONS ){
    int iColumn;
    while( readPosition(pReader, &iColumn)!=-1 ){}
  }
}

/* Skip over a docid, including its position list if the doclist has
 * positions. */
static void skipDocument(DocListReader *pReader){
  readDocid(pReader);
  skipPositionList(pReader);
}

/* Skip past all docids which are less than [iDocid].  Returns 1 if a docid
 * matching [iDocid] was found.  */
static int skipToDocid(DocListReader *pReader, sqlite_int64 iDocid){
  sqlite_int64 d = 0;
  while( !atEnd(pReader) && (d=peekDocid(pReader))<iDocid ){
    skipDocument(pReader);
  }
  return !atEnd(pReader) && d==iDocid;
}

/* Return the first document in a document list.
*/
static sqlite_int64 firstDocid(DocList *d){
  DocListReader r;
  readerInit(&r, d);
  return readDocid(&r);
}

#ifdef SQLITE_DEBUG
/*
** This routine is used for debugging purpose only.
**
** Write the content of a doclist to standard output.
*/
static void printDoclist(DocList *p){
  DocListReader r;
  const char *zSep = "";

  readerInit(&r, p);
  while( !atEnd(&r) ){
    sqlite_int64 docid = readDocid(&r);
    if( docid==0 ){
      skipPositionList(&r);
      continue;
    }
    printf("%s%lld", zSep, docid);
    zSep =  ",";
    if( p->iType>=DL_POSITIONS ){
      int iPos, iCol;
      const char *zDiv = "";
      printf("(");
      while( (iPos = readPosition(&r, &iCol))>=0 ){
        printf("%s%d:%d", zDiv, iCol, iPos);
        zDiv = ":";
      }
      printf(")");
    }
  }
  printf("\n");
  fflush(stdout);
}
#endif /* SQLITE_DEBUG */

/* Trim the given doclist to contain only positions in column
 * [iRestrictColumn]. */
static void docListRestrictColumn(DocList *in, int iRestrictColumn){
  DocListReader r;
  DocList out;

  assert( in->iType>=DL_POSITIONS );
  readerInit(&r, in);
  docListInit(&out, DL_POSITIONS, NULL, 0);

  while( !atEnd(&r) ){
    sqlite_int64 iDocid = readDocid(&r);
    int iPos, iColumn;

    docListAddDocid(&out, iDocid);
    while( (iPos = readPosition(&r, &iColumn)) != -1 ){
      if( iColumn==iRestrictColumn ){
        docListAddPos(&out, iColumn, iPos);
      }
    }
  }

  docListDestroy(in);
  *in = out;
}

/* Trim the given doclist by discarding any docids without any remaining
 * positions. */
static void docListDiscardEmpty(DocList *in) {
  DocListReader r;
  DocList out;

  /* TODO: It would be nice to implement this operation in place; that
   * could save a significant amount of memory in queries with long doclists. */
  assert( in->iType>=DL_POSITIONS );
  readerInit(&r, in);
  docListInit(&out, DL_POSITIONS, NULL, 0);

  while( !atEnd(&r) ){
    sqlite_int64 iDocid = readDocid(&r);
    int match = 0;
    int iPos, iColumn;
    while( (iPos = readPosition(&r, &iColumn)) != -1 ){
      if( !match ){
        docListAddDocid(&out, iDocid);
        match = 1;
      }
      docListAddPos(&out, iColumn, iPos);
    }
  }

  docListDestroy(in);
  *in = out;
}

/* Helper function for docListUpdate() and docListAccumulate().
** Splices a doclist element into the doclist represented by r,
** leaving r pointing after the newly spliced element.
*/
static void docListSpliceElement(DocListReader *r, sqlite_int64 iDocid,
                                 const char *pSource, int nSource){
  DocList *d = r->pDoclist;
  char *pTarget;
  int nTarget, found;

  found = skipToDocid(r, iDocid);

  /* Describe slice in d to place pSource/nSource. */
  pTarget = r->p;
  if( found ){
    skipDocument(r);
    nTarget = r->p-pTarget;
  }else{
    nTarget = 0;
  }

  /* The sense of the following is that there are three possibilities.
  ** If nTarget==nSource, we should not move any memory nor realloc.
  ** If nTarget>nSource, trim target and realloc.
  ** If nTarget<nSource, realloc then expand target.
  */
  if( nTarget>nSource ){
    memmove(pTarget+nSource, pTarget+nTarget, docListEnd(d)-(pTarget+nTarget));
  }
  if( nTarget!=nSource ){
    int iDoclist = pTarget-d->pData;
    d->pData = realloc(d->pData, d->nData+nSource-nTarget);
    pTarget = d->pData+iDoclist;
  }
  if( nTarget<nSource ){
    memmove(pTarget+nSource, pTarget+nTarget, docListEnd(d)-(pTarget+nTarget));
  }

  memcpy(pTarget, pSource, nSource);
  d->nData += nSource-nTarget;
  r->p = pTarget+nSource;
}

/* Insert/update pUpdate into the doclist. */
static void docListUpdate(DocList *d, DocList *pUpdate){
  DocListReader reader;

  assert( d!=NULL && pUpdate!=NULL );
  assert( d->iType==pUpdate->iType);

  readerInit(&reader, d);
  docListSpliceElement(&reader, firstDocid(pUpdate),
                       pUpdate->pData, pUpdate->nData);
}

/* Propagate elements from pUpdate to pAcc, overwriting elements with
** matching docids.
*/
static void docListAccumulate(DocList *pAcc, DocList *pUpdate){
  DocListReader accReader, updateReader;

  /* Handle edge cases where one doclist is empty. */
  assert( pAcc!=NULL );
  if( pUpdate==NULL || pUpdate->nData==0 ) return;
  if( pAcc->nData==0 ){
    pAcc->pData = malloc(pUpdate->nData);
    memcpy(pAcc->pData, pUpdate->pData, pUpdate->nData);
    pAcc->nData = pUpdate->nData;
    return;
  }

  readerInit(&accReader, pAcc);
  readerInit(&updateReader, pUpdate);

  while( !atEnd(&updateReader) ){
    char *pSource = updateReader.p;
    sqlite_int64 iDocid = readDocid(&updateReader);
    skipPositionList(&updateReader);
    docListSpliceElement(&accReader, iDocid, pSource, updateReader.p-pSource);
  }
}

/*
** Read the next docid off of pIn.  Return 0 if we reach the end.
*
* TODO: This assumes that docids are never 0, but they may actually be 0 since
* users can choose docids when inserting into a full-text table.  Fix this.
*/
static sqlite_int64 nextDocid(DocListReader *pIn){
  skipPositionList(pIn);
  return atEnd(pIn) ? 0 : readDocid(pIn);
}

/*
** pLeft and pRight are two DocListReaders that are pointing to
** positions lists of the same document: iDocid. 
**
** If there are no instances in pLeft or pRight where the position
** of pLeft is one less than the position of pRight, then this
** routine adds nothing to pOut.
**
** If there are one or more instances where positions from pLeft
** are exactly one less than positions from pRight, then add a new
** document record to pOut.  If pOut wants to hold positions, then
** include the positions from pRight that are one more than a
** position in pLeft.  In other words:  pRight.iPos==pLeft.iPos+1.
**
** pLeft and pRight are left pointing at the next document record.
*/
static void mergePosList(
  DocListReader *pLeft,    /* Left position list */
  DocListReader *pRight,   /* Right position list */
  sqlite_int64 iDocid,     /* The docid from pLeft and pRight */
  DocList *pOut            /* Write the merged document record here */
){
  int iLeftCol, iLeftPos = readPosition(pLeft, &iLeftCol);
  int iRightCol, iRightPos = readPosition(pRight, &iRightCol);
  int match = 0;

  /* Loop until we've reached the end of both position lists. */
  while( iLeftPos!=-1 && iRightPos!=-1 ){
    if( iLeftCol==iRightCol && iLeftPos+1==iRightPos ){
      if( !match ){
        docListAddDocid(pOut, iDocid);
        match = 1;
      }
      if( pOut->iType>=DL_POSITIONS ){
        docListAddPos(pOut, iRightCol, iRightPos);
      }
      iLeftPos = readPosition(pLeft, &iLeftCol);
      iRightPos = readPosition(pRight, &iRightCol);
    }else if( iRightCol<iLeftCol ||
              (iRightCol==iLeftCol && iRightPos<iLeftPos+1) ){
      iRightPos = readPosition(pRight, &iRightCol);
    }else{
      iLeftPos = readPosition(pLeft, &iLeftCol);
    }
  }
  if( iLeftPos>=0 ) skipPositionList(pLeft);
  if( iRightPos>=0 ) skipPositionList(pRight);
}

/* We have two doclists:  pLeft and pRight.
** Write the phrase intersection of these two doclists into pOut.
**
** A phrase intersection means that two documents only match
** if pLeft.iPos+1==pRight.iPos.
**
** The output pOut may or may not contain positions.  If pOut
** does contain positions, they are the positions of pRight.
*/
static void docListPhraseMerge(
  DocList *pLeft,    /* Doclist resulting from the words on the left */
  DocList *pRight,   /* Doclist for the next word to the right */
  DocList *pOut      /* Write the combined doclist here */
){
  DocListReader left, right;
  sqlite_int64 docidLeft, docidRight;

  readerInit(&left, pLeft);
  readerInit(&right, pRight);
  docidLeft = nextDocid(&left);
  docidRight = nextDocid(&right);

  while( docidLeft>0 && docidRight>0 ){
    if( docidLeft<docidRight ){
      docidLeft = nextDocid(&left);
    }else if( docidRight<docidLeft ){
      docidRight = nextDocid(&right);
    }else{
      mergePosList(&left, &right, docidLeft, pOut);
      docidLeft = nextDocid(&left);
      docidRight = nextDocid(&right);
    }
  }
}

/* We have two doclists:  pLeft and pRight.
** Write the intersection of these two doclists into pOut.
** Only docids are matched.  Position information is ignored.
**
** The output pOut never holds positions.
*/
static void docListAndMerge(
  DocList *pLeft,    /* Doclist resulting from the words on the left */
  DocList *pRight,   /* Doclist for the next word to the right */
  DocList *pOut      /* Write the combined doclist here */
){
  DocListReader left, right;
  sqlite_int64 docidLeft, docidRight;

  assert( pOut->iType<DL_POSITIONS );

  readerInit(&left, pLeft);
  readerInit(&right, pRight);
  docidLeft = nextDocid(&left);
  docidRight = nextDocid(&right);

  while( docidLeft>0 && docidRight>0 ){
    if( docidLeft<docidRight ){
      docidLeft = nextDocid(&left);
    }else if( docidRight<docidLeft ){
      docidRight = nextDocid(&right);
    }else{
      docListAddDocid(pOut, docidLeft);
      docidLeft = nextDocid(&left);
      docidRight = nextDocid(&right);
    }
  }
}

/* We have two doclists:  pLeft and pRight.
** Write the union of these two doclists into pOut.
** Only docids are matched.  Position information is ignored.
**
** The output pOut never holds positions.
*/
static void docListOrMerge(
  DocList *pLeft,    /* Doclist resulting from the words on the left */
  DocList *pRight,   /* Doclist for the next word to the right */
  DocList *pOut      /* Write the combined doclist here */
){
  DocListReader left, right;
  sqlite_int64 docidLeft, docidRight, priorLeft;

  readerInit(&left, pLeft);
  readerInit(&right, pRight);
  docidLeft = nextDocid(&left);
  docidRight = nextDocid(&right);

  while( docidLeft>0 && docidRight>0 ){
    if( docidLeft<=docidRight ){
      docListAddDocid(pOut, docidLeft);
    }else{
      docListAddDocid(pOut, docidRight);
    }
    priorLeft = docidLeft;
    if( docidLeft<=docidRight ){
      docidLeft = nextDocid(&left);
    }
    if( docidRight>0 && docidRight<=priorLeft ){
      docidRight = nextDocid(&right);
    }
  }
  while( docidLeft>0 ){
    docListAddDocid(pOut, docidLeft);
    docidLeft = nextDocid(&left);
  }
  while( docidRight>0 ){
    docListAddDocid(pOut, docidRight);
    docidRight = nextDocid(&right);
  }
}

/* We have two doclists:  pLeft and pRight.
** Write into pOut all documents that occur in pLeft but not
** in pRight.
**
** Only docids are matched.  Position information is ignored.
**
** The output pOut never holds positions.
*/
static void docListExceptMerge(
  DocList *pLeft,    /* Doclist resulting from the words on the left */
  DocList *pRight,   /* Doclist for the next word to the right */
  DocList *pOut      /* Write the combined doclist here */
){
  DocListReader left, right;
  sqlite_int64 docidLeft, docidRight, priorLeft;

  readerInit(&left, pLeft);
  readerInit(&right, pRight);
  docidLeft = nextDocid(&left);
  docidRight = nextDocid(&right);

  while( docidLeft>0 && docidRight>0 ){
    priorLeft = docidLeft;
    if( docidLeft<docidRight ){
      docListAddDocid(pOut, docidLeft);
    }
    if( docidLeft<=docidRight ){
      docidLeft = nextDocid(&left);
    }
    if( docidRight>0 && docidRight<=priorLeft ){
      docidRight = nextDocid(&right);
    }
  }
  while( docidLeft>0 ){
    docListAddDocid(pOut, docidLeft);
    docidLeft = nextDocid(&left);
  }
}

static char *string_dup_n(const char *s, int n){
  char *str = malloc(n + 1);
  memcpy(str, s, n);
  str[n] = '\0';
  return str;
}

/* Duplicate a string; the caller must free() the returned string.
 * (We don't use strdup() since it's not part of the standard C library and
 * may not be available everywhere.) */
static char *string_dup(const char *s){
  return string_dup_n(s, strlen(s));
}