CARL Source
Full-text search
Ergonomics > Full-text search

The full-text search can be accessed in the top panel of the application.

The availability of full-text search depends on your installation.
It will only be accessible if an Elasticsearch cluster® has been deployed in your environment.

View

Full-text search:

When the search is performed, the results appear in a suggestion list containing the following information:

 

Function-type result

→ Clicking on a line in the suggestion list opens the function search results.

 

Entity-type result

→ Clicking on a line in the list of suggestions opens the entity's details.

 

Configuration

By default, all the entities declared in the elasticIndexing.xml file are indexed by the code and name. These are the entities containing an associated detail screen.

It is possible to index other attributes from the dictionary configuration.

 

Sometimes the data of an object are not returned by the full-text search even though the object is specified as "Indexed".
Using a few actions, it is possible to check whether a table has been indexed and, if not, to relaunch indexing.

Prerequisites

  1. On the Elasticsearch® host sever, the on-line command tool cURL must be available.
  2. In CARL Source, an automatic job using the class CSJOBINDEXER must be present.


Operations

Connect to a server hosting one of the nodes of the Elasticsearch® cluster of your environment.

  1. To view the list of indexes already referenced, use a URL browser to view the following: http://localhost:port/_cat/indices
    → Replace port with the port used by Elasticsearch® (by default, this port 9200, but check the configuration in the declarations of CARL Source Admin).

  2. If the searched table appears on this list, you can clean all of the indexes by running the following command on a terminal: curl -X DELETE http://localhost:port/_all
    → Again, be sure to adapt the value corresponding to port.

  3. In CARL Source, run the automatic job using the class CSJOBINDEXER.
    → Si the job completes correctly, the indexing must take into account the data that had not been returned by the search until that point.