Elasticsearch is a NoSQL Database focused on search capabilities to act as a search engine. The Shopware implementation of Elasticsearch provides an integrated way to improve the performance of product and category searches. To use Elasticsearch for your shop take a look at our elasticsearch guide
Elasticsearch is only used in searches that are explicitly defined. This is by default set to the
ProductSuggestRoute. To use elasticsearch on your own searches make sure to add the elasticsearch aware state to your criteria.
$criteria = new \Shopware\Core\Framework\DataAbstractionLayer\Search\Criteria();$context = \Shopware\Core\Framework\Context::createDefaultContext();// Enables elasticsearch for this search$context->addState(\Shopware\Core\Framework\Context::STATE_ELASTICSEARCH_AWARE);$repository->search($criteria, $context);
To provide elasticsearch for an entity a corresponding
ElasticsearchDefinition needs to be added. Currently, Shopware has such a definition for the product entity called
ProductElasticsearchDefinition. This definition defines the fields which are provided to elasticsearch and how they are aggregated.
ElasticsearchEntitySearcher decorates the
EntitySearcher to map the entity search to the elasticsearch structure. The
ElasticsearchEntitySearcher returns an
IdSearchResult hydrated by the
ElasticsearchEntitySearchHydrator as the
EntitySearcher does and this result is used to read the found ids from the database.
ElasticsearchEntityAggregator does the same as the
ElasticsearchEntitySearcher for aggregations.
CriteriaParser parses the criteria to an elasticsearch specific notation.
The product search has a special
ProductSearchBuilder in the core and so has the elasticsearch extension a corresponding extension for the
ProductSearchBuilder. This extension matches the queries of the core
ProductSearchBuilder to the elasticsearch notation.
ProductUpdater listens to the
ProductIndexerEvent and triggers the
ElasticsearchIndexer on changes to a
es:index:cleanup deletes outdated elasticsearch indexes. The parameter
-f will skip the confirmation.
es:create:alias refreshes the current elasticsearch index and sets the alias to the index name without the timestamp (which will make this index the active index). This will happen automatically when a new index is published, so this command can force the alias creation for testing purposes or if something went wrong.
es:index reindexes all configured entities to elasticsearch.
es:reset resets all active indices and clears the queue. This command should be used only if an index is corrupted or needs to be setup completely from scratch.
es:status returns the status of all current elasticsearch indices.
es:test:analyzer runs an elasticsearch analyzer on your indices. For more details on elasticsearch analyzers take a look on the elasticsearch reference (external link).
To customize the Elasticsearch integration or add own fields and entities refer to the elasticsearch extension guide