Ignore:
Timestamp:
2008-08-21T15:13:41+12:00 (16 years ago)
Author:
mdewsnip
Message:

Changed the Lucene code to use the Greenstone document OIDs directly, instead of creating its own numeric IDs and then mapping them to the Greenstone OIDs in the GDBM file. As well as being simpler and more space and speed efficient (the mapping no longer needs to be stored in the GDBM file, and no lookup needs to be done for each search result), this is another important step along the road to true incremental building.

File:
1 edited

Legend:

Unmodified
Added
Removed
  • gsdl/trunk/runtime-src/src/colservr/expat_resultset.cpp

    r16915 r16947  
    9191
    9292    char* id = get_attribute(attributes,"id");
    93     if (id!=NULL) {
    94       int id_num = atoi(id);
    95    
     93    if (id != NULL)
     94    {
    9695      docresultclass doc;
    9796      doc.clear();
    98       doc.docnum = id_num;
     97      doc.docid = id;
    9998      doc.docweight = qrpack_ptr->match_count;
    10099
     
    105104      }
    106105
    107       queryresult_ptr->docs.docset[doc.docnum] = doc;
    108       queryresult_ptr->docs.docorder.push_back(doc.docnum);
     106      queryresult_ptr->docs.docset[doc.docid] = doc;
     107      queryresult_ptr->docs.docorder.push_back(doc.docid);
    109108      ++qrpack_ptr->match_count;
    110109
Note: See TracChangeset for help on using the changeset viewer.