Exemple #1
0
 public static void indexSchemas()
     throws IOException, SAXException, ParserConfigurationException, SolrServerException {
   List<SecureSchema> schemas = SecureSchema.findAll(NoSql.em());
   for (SecureSchema schema : schemas) {
     SearchUtils.indexSchema(schema);
   }
 }
Exemple #2
0
  @Util
  public static void reindex()
      throws IOException, SAXException, ParserConfigurationException, SolrServerException {

    deleteExistingCores();
    indexSchemas();

    Cursor<KeyValue<SecureTable>> tablesCursor = SecureTable.findAllCursor(NoSql.em());
    int i = 0;
    // very important to ignore any error and continue indexing.  If an index has gotten
    // corrupted (which happens...) this will throw com.alvazan.orm.api.exc.RowNotFoundException
    // and will kill the entire reindex.
    long docsindexed = 0;
    long startTime = System.currentTimeMillis() - 1;
    Collection<SolrInputDocument> solrDocs = new ArrayList<SolrInputDocument>();
    while (tablesCursor.next()) {
      if (++i % 200 == 0) NoSql.em().clear();
      KeyValue<SecureTable> kv = tablesCursor.getCurrent();

      try {
        if (kv.getValue() == null) continue;
      } catch (RowNotFoundException rnfe) {
        if (log.isInfoEnabled())
          log.error(
              "got a corrupt index while reindexing, ignoring the error and continuing with indexing of other data.");
        // rnfe.printStackTrace();
        continue;
      }
      SecureTable table = kv.getValue();
      DboTableMeta meta = table.getTableMeta();

      SearchUtils.indexTable(table, meta, solrDocs);

      if (table.isSearchable()) {
        log.info("found a searchable table " + table.getName() + " indexing it.");
        String sql = "select c from " + table.getTableName() + " as c";
        Collection<SolrInputDocument> tablesolrDocs = new ArrayList<SolrInputDocument>();

        try {
          QueryResult result =
              NoSql.em().getTypedSession().createQueryCursor(sql, SqlPullProcessor.BATCH_SIZE);
          Iterator<List<TypedRow>> cursor = result.getAllViewsIter().iterator();

          while (true) {
            // I hate this, but cursor.hasNext() can throw an exception which means we need to skip
            // over
            // that item but continue on with the cursor till it runs out:
            List<TypedRow> typedRows = getNext(cursor);
            if (typedRows == null) break;
            for (TypedRow prow : typedRows) {
              SearchPosting.addSolrDataDoc(prow, table, tablesolrDocs);
            }
            if (tablesolrDocs.size() > REINDEX_BATCH_SIZE) {
              docsindexed += solrDocs.size();
              System.out.println(
                  "hit solr doc batch size in a searchable table, "
                      + docsindexed
                      + " docs so far, "
                      + (System.currentTimeMillis() - startTime)
                      + " millis elapsed "
                      + (docsindexed / ((System.currentTimeMillis() - startTime) / 1000))
                      + " docs per sec.");
              SearchPosting.saveSolr("reindex", tablesolrDocs, null);
              tablesolrDocs = new ArrayList<SolrInputDocument>();
            }
          }
          SearchPosting.saveSolr("reindex", tablesolrDocs, null);
          docsindexed += solrDocs.size();
        } catch (Exception e) {
          System.out.println(
              "got an exception while indexing a searchable table with the query (probably a corrupt index in playorm):");
          System.out.println(sql);
          // e.printStackTrace();
        }
      }
      if (solrDocs.size() > REINDEX_BATCH_SIZE) {
        docsindexed += solrDocs.size();
        System.out.println(
            "hit solr doc batch size in metadata, "
                + docsindexed
                + " docs so far, "
                + (System.currentTimeMillis() - startTime)
                + " millis elapsed "
                + (docsindexed / ((System.currentTimeMillis() - startTime) / 1000))
                + " docs per sec.");
        SearchPosting.saveSolr("reindex", solrDocs, "databusmeta");
        solrDocs = new ArrayList<SolrInputDocument>();
      }
    }
    if (solrDocs.size() > 0) {
      docsindexed += solrDocs.size();
      System.out.println(
          "hit solr doc batch size during finalization, "
              + docsindexed
              + " docs so far, "
              + (System.currentTimeMillis() - startTime)
              + " millis elapsed "
              + (docsindexed / ((System.currentTimeMillis() - startTime) / 1000))
              + " docs per sec.");
      SearchPosting.saveSolr("reindex", solrDocs, "databusmeta");
      solrDocs = new ArrayList<SolrInputDocument>();
    }
  }