Halyard

What it does

Bulk data tools

Its command-line and MapReduce tools support bulk loading and exporting RDF data, dataset statistics, SPARQL updates, and deletion of large sets of triples or named graphs.checked 8 Oct 2026

Bulk loading

Halyard Bulk Load uses MapReduce to load RDF data from HDFS into HBase, and the documented input formats include Turtle, RDF/XML, N-Triples, TriG, N-Quads, and JSON-LD.checked 8 Oct 2026

Data deletion behavior

Deleting a repository configuration does not delete its Halyard dataset, while clearing the repository or deleting statements affects the shared data for all users.checked 8 Oct 2026

Data management

Command-line tools support SPARQL Update, parallel bulk updates, export, and bulk deletion of triples or named graphs.checked 8 Oct 2026

Data storage

Halyard datasets are stored as HBase tables, and multiple repositories can point to one shared dataset.checked 8 Oct 2026

Deployment

The maker documents deployment on Amazon Elastic MapReduce or on an Apache Hadoop cluster node configured with an Apache HBase client.checked 8 Oct 2026

Endpoint

Halyard Endpoint is a command-line application for launching a SPARQL endpoint to serve queries.checked 8 Oct 2026

Experimental feature

Halyard Summary is described as an experimental MapReduce application that calculates an approximate dataset summary and exports it to a file.checked 8 Oct 2026

Federation limit

Federated queries can span Halyard datasets using SPARQL SERVICE, but external SPARQL endpoint service types are not recognized.checked 8 Oct 2026

Implementation

Halyard is written in Java and is based on Eclipse RDF4J and Apache HBase.checked 8 Oct 2026

Not bundled

Apache Hadoop and Apache HBase components are not bundled with Halyard and must be available in the runtime environment.checked 8 Oct 2026

Product

Halyard is a horizontally scalable RDF store with named graph support, designed for very large semantic data models and SPARQL 1.1 queries over Linked Data snapshots.checked 8 Oct 2026

Purpose

Halyard is a horizontally scalable RDF store with named graph support for storing and integrating large semantic data models and querying linked data snapshots with SPARQL 1.1.checked 8 Oct 2026

Querying

The included RDF4J Console can evaluate SPARQL queries and connect to local or remote repository endpoints.checked 8 Oct 2026

Query interface

Halyard supports SPARQL access through RDF4J Console, RDF4J Workbench, and RDF4J Server SPARQL endpoint REST APIs.checked 8 Oct 2026

RDF formats

Bulk Load accepts RDF4J RIO-supported formats including N-Triples, RDF/XML, Turtle, TriG, N-Quads, and JSON-LD, with Hadoop-supported compression codecs.checked 8 Oct 2026

Runtime requirements

The getting-started page lists Apache Hadoop 2.5.1 or later, Apache HBase 1.1.2 or later, and Java 8 Runtime as requirements.checked 8 Oct 2026

Search integration

An optional supplementary Elasticsearch index can index dataset literals to provide more advanced text search features.checked 8 Oct 2026

Security behavior

On a secured cluster, the RDF4J Console instructions say to authenticate with kinit credentials before use.checked 8 Oct 2026

Statistics

Halyard Stats calculates dataset statistics and stores them in a named graph or exports them to a file.checked 8 Oct 2026

Storage modes

For Amazon EMR deployments, HBase can use HDFS or S3 storage.checked 8 Oct 2026

Support

The home page links to documentation, open GitHub issues, a discussion group, and lists Adam Sotona as the author contact.checked 8 Oct 2026

Technology

Halyard is written in Java and built on Eclipse RDF4J and Apache HBase.checked 8 Oct 2026