How-to guides
Practical, specific walk-throughs for working with the ociforge Ocient® sandbox — from authentication to writing queries.
The smallest Ocient system
Ocient is built for petabyte, trillion-row hyperscale — yet ociforge runs the same software on the smallest possible footprint: three nodes, one NVMe drive each. How Ocient scales horizontally across loader, SQL, and foundation nodes.
Setting up Ocient OAuth/OIDC SSO with Auth0
Put an Ocient System behind OIDC single sign-on: an SSO-only connectivity pool, configuring SSO once on the system database for every database, and separate device-grant vs machine-to-machine applications.
Interesting queries for the ociforge sandbox
Example SQL over 550M+ rows of NYC taxi, NOAA weather, and GDELT world-event data — joins, time-series pruning, geospatial point-in-polygon, and approximate analytics on Ocient.
How the sandbox data was loaded: pipelines
How ociforge's 550M+ rows were loaded with Ocient data pipelines — anonymous public S3, Parquet from the filesystem, tab-delimited GDELT, and load-time transformations including geospatial POINT and POLYGON construction.
What the database remembers: sys.completed_queries
Every Ocient query is recorded in sys.completed_queries with 72 columns of timing, cost, and workload-management detail. How to read it, and how it ties to service classes and query priority.