Posted On: Sep 3, 2021

Amazon Neptune now supports searching on new data types, such as numbers and dates, in addition to strings when using the full-text search integration with Elasticsearch. This improvement allows Neptune customers to replicate non-string values into an Elasticsearch cluster, such as provided by the Amazon Elasticsearch Service, to run Gremlin or SPARQL queries searching on these values.

Customers asked for more ways to search and filter on graph data using the Neptune full-text search integration with Elasticsearch. Now you can access Elasticsearch’s built-in indexing by text, longs, doubles, booleans, and dates when querying their Neptune graphs. Non-string indexing is enabled by default and available starting in engine release version You can use the quick-start to set up the full-text search integration for the first time, or update the existing full-text search integration by performing one-time re-indexing. Gremlin users can use the withSideEffect step and pass the Elasticsearch endpoint, search pattern, and field information. Similarly, SPARQL users can use the SERVICE keyword to federate queries to Elasticsearch.

To learn more about this integration, including sample search queries for non-string values, read the Amazon Neptune User Guide. Customers can use the Amazon Neptune integration with Elasticsearch clusters in all regions where Neptune is available. There are no additional charges for using this feature. For more information on pricing and region availability, refer to the Neptune pricing page and AWS Region Table.