Important

You are browsing the documentation for version 4.2 of OroCommerce, OroCRM and OroPlatform, which is no longer maintained. Read version 5.1 (the latest LTS version) of the Oro documentation to get up-to-date information.

See our Release Process documentation for more information on the currently supported and upcoming releases.

ORM Search Engine

ORM engine is a default search engine provided out of the box by WebsiteSearchBundle. This engine uses Doctrine ORM entities as a storage, so all data is stored in a relational DBMS. ORM search index shares the same database with an application. To perform fulltext search index uses DBMS fulltext search index.

ORM Data Storage

ORM storage uses EAV (entity-attribute-value) model to store multiple fields of the same entity. There are five ORM entities in the WebsiteSearchBundle to store search index information:

  • Oro\Bundle\WebsiteSearchBundle\Entity\Item (table oro_website_search_item) - main entry point entity, stores general information about an entity (name, ID, alias) and includes relations to all other search entities;

  • Oro\Bundle\WebsiteSearchBundle\Entity\IndexText (table oro_website_search_text) - stores values of fields of the text type , contains fulltext search index;

  • Oro\Bundle\WebsiteSearchBundle\Entity\IndexInteger (table oro_website_search_integer) - stores values of integer fields;

  • Oro\Bundle\WebsiteSearchBundle\Entity\IndexDecimal (table oro_website_search_decimal) - stores values of decimal fields;

  • Oro\Bundle\WebsiteSearchBundle\Entity\IndexDatetime (table oro_website_search_datetime) - stores values of datatime fields.

Each of these entities has its own table in the database. Four type-specific tables have relation to main entity table. Entities from the WebsiteSearchBundle use separate entity manager – search – and, as a consequence, separate connection to the database.

ORM engine supports two DBMSes - Mysql and PostgreSQL. Each of these DBMSese has it’s own driver class that encapsulates interaction and provides common interface to execute queries. Mysql driver is stored in Oro\Bundle\WebsiteSearchBundle\Engine\ORM\Driver\PdoMysql, PostgreSQL driver is stored in Oro\Bundle\WebsiteSearchBundle\Engine\ORM\Driver\PdoPgsql - both of these drivers extend similar drivers from SearchBundle and use the same approach to work with search index.

The ORM engine has some advantages and disadvantages. One of the big advantages is that engine shares the same database with an application, so it can be backed up with the main data, and there is no need to set up separate search engine like Elasticsearch or Sphinx. With relational DBMSes, indexation happens faster.

On the other hand, the search via ORM fulltext search index is not that fast, especially if there are many enitites. One more disadvantage is the usage of EAV. Though it is very flexible, the database query execution might be quite heavy and memory consuming.

ORM Indexation

An ORM indexer is represented by the Oro\Bundle\WebsiteSearchBundle\Engine\ORM\OrmIndexer class and almost all method calls proxy methods from an appropriate ORM search driver.

The only interesting part in this indexer is alias renaming. When developer requests reindexation of some website (or full reindexation), the data is not removed from ORM immediately. Instead, a new temporary alias is used to index new data.

Then, after all new data is persisted, the old data with its permanent alias is dropped and the temporary alias is renamed to the permanent one. With this approach, search index always contains some data, and user is able to use search during the indexation.