Example #1
0
 @SuppressWarnings("unchecked")
 private void addFields(
     DataConfig.Entity entity, DocWrapper doc, Map<String, Object> arow, VariableResolver vr) {
   for (Map.Entry<String, Object> entry : arow.entrySet()) {
     String key = entry.getKey();
     Object value = entry.getValue();
     if (value == null) continue;
     if (key.startsWith("$")) continue;
     List<DataConfig.Field> field = entity.colNameVsField.get(key);
     if (field == null && dataImporter.getSchema() != null) {
       // This can be a dynamic field or a field which does not have an entry in data-config ( an
       // implicit field)
       SchemaField sf = dataImporter.getSchema().getFieldOrNull(key);
       if (sf == null) {
         sf = dataImporter.getConfig().lowerNameVsSchemaField.get(key.toLowerCase(Locale.ENGLISH));
       }
       if (sf != null) {
         addFieldToDoc(entry.getValue(), sf.getName(), 1.0f, sf.multiValued(), doc);
       }
       // else do nothing. if we add it it may fail
     } else {
       if (field != null) {
         for (DataConfig.Field f : field) {
           String name = f.getName();
           if (f.dynamicName) {
             name = vr.replaceTokens(name);
           }
           if (f.toWrite) addFieldToDoc(entry.getValue(), name, f.boost, f.multiValued, doc);
         }
       }
     }
   }
 }
Example #2
0
 private void invokeEventListener(String className) {
   try {
     EventListener listener =
         (EventListener) loadClass(className, dataImporter.getCore()).newInstance();
     notifyListener(listener);
   } catch (Exception e) {
     wrapAndThrow(SEVERE, e, "Unable to load class : " + className);
   }
 }
Example #3
0
 private void notifyListener(EventListener listener) {
   String currentProcess;
   if (dataImporter.getStatus() == DataImporter.Status.RUNNING_DELTA_DUMP) {
     currentProcess = Context.DELTA_DUMP;
   } else {
     currentProcess = Context.FULL_DUMP;
   }
   listener.onEvent(
       new ContextImpl(null, getVariableResolver(), null, currentProcess, session, null, this));
 }
Example #4
0
 @SuppressWarnings("unchecked")
 private void doFullDump() {
   addStatusMessage("Full Dump Started");
   if (dataImporter.getConfig().isMultiThreaded && !verboseDebug) {
     try {
       LOG.info("running multithreaded full-import");
       new EntityRunner(root, null).run(null, Context.FULL_DUMP, null);
     } catch (Exception e) {
       LOG.error("error in import", e);
     }
   } else {
     buildDocument(getVariableResolver(), null, null, root, true, null);
   }
 }
Example #5
0
 public VariableResolverImpl getVariableResolver() {
   try {
     VariableResolverImpl resolver = null;
     if (dataImporter != null && dataImporter.getCore() != null) {
       resolver =
           new VariableResolverImpl(
               dataImporter.getCore().getResourceLoader().getCoreProperties());
     } else resolver = new VariableResolverImpl();
     Map<String, Object> indexerNamespace = new HashMap<String, Object>();
     if (persistedProperties.getProperty(LAST_INDEX_TIME) != null) {
       indexerNamespace.put(LAST_INDEX_TIME, persistedProperties.getProperty(LAST_INDEX_TIME));
     } else {
       // set epoch
       indexerNamespace.put(LAST_INDEX_TIME, DataImporter.DATE_TIME_FORMAT.get().format(EPOCH));
     }
     indexerNamespace.put(INDEX_START_TIME, dataImporter.getIndexStartTime());
     indexerNamespace.put("request", requestParameters.requestParams);
     indexerNamespace.put("functions", functionsNamespace);
     for (DataConfig.Entity entity : dataImporter.getConfig().document.entities) {
       String key = entity.name + "." + SolrWriter.LAST_INDEX_KEY;
       String lastIndex = persistedProperties.getProperty(key);
       if (lastIndex != null) {
         indexerNamespace.put(key, lastIndex);
       } else {
         indexerNamespace.put(key, DataImporter.DATE_TIME_FORMAT.get().format(EPOCH));
       }
     }
     resolver.addNamespace(DataConfig.IMPORTER_NS_SHORT, indexerNamespace);
     resolver.addNamespace(DataConfig.IMPORTER_NS, indexerNamespace);
     return resolver;
   } catch (Exception e) {
     wrapAndThrow(SEVERE, e);
     // unreachable statement
     return null;
   }
 }
Example #6
0
 private EntityProcessorWrapper getEntityProcessor(DataConfig.Entity entity) {
   if (entity.processor != null) return entity.processor;
   EntityProcessor entityProcessor = null;
   if (entity.proc == null) {
     entityProcessor = new SqlEntityProcessor();
   } else {
     try {
       entityProcessor =
           (EntityProcessor) loadClass(entity.proc, dataImporter.getCore()).newInstance();
     } catch (Exception e) {
       wrapAndThrow(
           SEVERE, e, "Unable to load EntityProcessor implementation for entity:" + entity.name);
     }
   }
   return entity.processor = new EntityProcessorWrapper(entityProcessor, this);
 }
Example #7
0
  @SuppressWarnings("unchecked")
  public void execute() {
    dataImporter.store(DataImporter.STATUS_MSGS, statusMessages);
    document = dataImporter.getConfig().document;
    final AtomicLong startTime = new AtomicLong(System.currentTimeMillis());
    statusMessages.put(
        TIME_ELAPSED,
        new Object() {
          public String toString() {
            return getTimeElapsedSince(startTime.get());
          }
        });

    statusMessages.put(DataImporter.MSG.TOTAL_QUERIES_EXECUTED, importStatistics.queryCount);
    statusMessages.put(DataImporter.MSG.TOTAL_ROWS_EXECUTED, importStatistics.rowsCount);
    statusMessages.put(DataImporter.MSG.TOTAL_DOC_PROCESSED, importStatistics.docCount);
    statusMessages.put(DataImporter.MSG.TOTAL_DOCS_SKIPPED, importStatistics.skipDocCount);

    List<String> entities = requestParameters.entities;

    // Trigger onImportStart
    if (document.onImportStart != null) {
      invokeEventListener(document.onImportStart);
    }
    AtomicBoolean fullCleanDone = new AtomicBoolean(false);
    // we must not do a delete of *:* multiple times if there are multiple root entities to be run
    Properties lastIndexTimeProps = new Properties();
    lastIndexTimeProps.setProperty(
        LAST_INDEX_KEY,
        DataImporter.DATE_TIME_FORMAT.get().format(dataImporter.getIndexStartTime()));
    for (DataConfig.Entity e : document.entities) {
      if (entities != null && !entities.contains(e.name)) continue;
      lastIndexTimeProps.setProperty(
          e.name + "." + LAST_INDEX_KEY, DataImporter.DATE_TIME_FORMAT.get().format(new Date()));
      root = e;
      String delQuery = e.allAttributes.get("preImportDeleteQuery");
      if (dataImporter.getStatus() == DataImporter.Status.RUNNING_DELTA_DUMP) {
        cleanByQuery(delQuery, fullCleanDone);
        doDelta();
        delQuery = e.allAttributes.get("postImportDeleteQuery");
        if (delQuery != null) {
          fullCleanDone.set(false);
          cleanByQuery(delQuery, fullCleanDone);
        }
      } else {
        cleanByQuery(delQuery, fullCleanDone);
        doFullDump();
        delQuery = e.allAttributes.get("postImportDeleteQuery");
        if (delQuery != null) {
          fullCleanDone.set(false);
          cleanByQuery(delQuery, fullCleanDone);
        }
      }
      statusMessages.remove(DataImporter.MSG.TOTAL_DOC_PROCESSED);
    }

    if (stop.get()) {
      // Dont commit if aborted using command=abort
      statusMessages.put("Aborted", DataImporter.DATE_TIME_FORMAT.get().format(new Date()));
      rollback();
    } else {
      // Do not commit unnecessarily if this is a delta-import and no documents were created or
      // deleted
      if (!requestParameters.clean) {
        if (importStatistics.docCount.get() > 0 || importStatistics.deletedDocCount.get() > 0) {
          finish(lastIndexTimeProps);
        }
      } else {
        // Finished operation normally, commit now
        finish(lastIndexTimeProps);
      }

      if (writer != null) {
        writer.finish();
      }

      if (document.onImportEnd != null) {
        invokeEventListener(document.onImportEnd);
      }
    }

    statusMessages.remove(TIME_ELAPSED);
    statusMessages.put(DataImporter.MSG.TOTAL_DOC_PROCESSED, "" + importStatistics.docCount.get());
    if (importStatistics.failedDocCount.get() > 0)
      statusMessages.put(
          DataImporter.MSG.TOTAL_FAILED_DOCS, "" + importStatistics.failedDocCount.get());

    statusMessages.put("Time taken ", getTimeElapsedSince(startTime.get()));
    LOG.info("Time taken = " + getTimeElapsedSince(startTime.get()));
  }