# AI use cases for Open Lakehouse Compute Fabric (jddevtime/open-lakehouse-compute-fabric) Repo: [jddevtime/open-lakehouse-compute-fabric](https://github.com/jddevtime/open-lakehouse-compute-fabric) | Use cases: 5 | Page: https://dedupedai.com/use-cases/jddevtime-open-lakehouse-compute-fabric/ Open Lakehouse Compute Fabric (OLCF) is an Apache 2.0 Python platform with a console, REST API, CLI, SDK and MCP endpoint for lakehouse work, plus a Spark compute fabric you run on your own machines and an extension for Microsoft Fabric notebooks. This file holds every use case of this repo. Each one starts with a "## " heading and ends with its URL. - Run a full lakehouse platform on one machine: https://dedupedai.com/use-cases/jddevtime-open-lakehouse-compute-fabric/run-full-lakehouse-on-one-machine/ - Find what depends on a table and who owns it: https://dedupedai.com/use-cases/jddevtime-open-lakehouse-compute-fabric/trace-table-lineage-and-owners/ - Let Claude Code browse the catalog and run read-only SQL: https://dedupedai.com/use-cases/jddevtime-open-lakehouse-compute-fabric/let-claude-code-query-the-catalog/ - Run a Fabric notebook on your own Spark hardware: https://dedupedai.com/use-cases/jddevtime-open-lakehouse-compute-fabric/run-fabric-notebook-on-own-spark/ - Build a Spark cluster from your own machines: https://dedupedai.com/use-cases/jddevtime-open-lakehouse-compute-fabric/build-spark-fabric-across-own-machines/ ## How to run a full lakehouse platform on one machine with Open Lakehouse Compute Fabric Repo: [jddevtime/open-lakehouse-compute-fabric](https://github.com/jddevtime/open-lakehouse-compute-fabric) | Difficulty: Intermediate | Time: 30 minutes | Basis: From the README Start the console, SQL, notebooks and sample Retail lakehouse on a laptop with Python only. **Who it is for:** A data analyst or developer who wants to try a lakehouse platform locally, with no cloud account and no Spark install. **The problem:** Setting up a lakehouse with catalog, SQL, notebooks and dashboards usually means several services and cloud accounts before any data can be queried. **The outcome:** A running OLCF console on one machine with SQL on DuckDB, Python notebooks, dashboards, pipelines and a sample Retail lakehouse. ### Steps 1. Install OLCF using the Install section of the README. 2. Run `olcf quickstart`. 3. Open the console the quickstart reports and open the sample Retail lakehouse. 4. Run a query in the SQL editor against the sample lakehouse and check that rows come back. ### Prompt to paste into your agent ```text Use the open-lakehouse-compute-fabric repo (jddevtime/open-lakehouse-compute-fabric). Install it from the README Install section, run `olcf quickstart`, and report the console address and whether the sample Retail lakehouse loads. Verify by running one SQL query against the sample lakehouse and returning the row count. ``` ### Caveat The README marks this release as 0.1.0 beta for evaluation on trusted infrastructure, so do not point it at production data yet. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/jddevtime-open-lakehouse-compute-fabric/run-full-lakehouse-on-one-machine/ ## How to find what depends on a table and who owns it with Open Lakehouse Compute Fabric Repo: [jddevtime/open-lakehouse-compute-fabric](https://github.com/jddevtime/open-lakehouse-compute-fabric) | Difficulty: Intermediate | Time: 30 minutes | Basis: From the README Use OLCF Intelligence to follow table lineage and see the confidence and gaps for each edge. **Who it is for:** A data steward who must answer what depends on a table before changing or retiring it. **The problem:** Before changing a table, a steward needs to know which notebooks, pipelines and reports read it, and whether those links are proven or only guessed. **The outcome:** A lineage view of upstream and downstream assets, with each edge marked verified, derived, inferred or unknown, and a list of gaps with reasons. ### Steps 1. Complete the quickstart so the platform has a catalog to scan. 2. Open the Intelligence section of the console, or run `olcf intelligence` from the CLI. 3. Search for the table and open its lineage view. 4. Review each edge's confidence level and the listed gaps before making a change. ### Prompt to paste into your agent ```text Use the open-lakehouse-compute-fabric repo (jddevtime/open-lakehouse-compute-fabric). Run `olcf intelligence` against the local installation, and for list its upstream and downstream assets with the confidence level of each edge. Verify that every edge has a confidence value and that gaps are listed with reasons, not guessed. ``` ### Caveat Lineage is only as complete as the scanned sources, and edges inferred by name are not proof of a real dependency. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/jddevtime-open-lakehouse-compute-fabric/trace-table-lineage-and-owners/ ## How to let Claude Code browse the catalog and run read-only SQL with Open Lakehouse Compute Fabric Repo: [jddevtime/open-lakehouse-compute-fabric](https://github.com/jddevtime/open-lakehouse-compute-fabric) | Difficulty: Intermediate | Time: 45 minutes | Basis: From the README Connect Claude Code to the OLCF MCP endpoint to explore tables, follow lineage and query safely. **Who it is for:** An analyst who wants an AI coding agent to answer questions about the lakehouse without giving it write access. **The problem:** Answering questions about tables and lineage means manual catalog browsing and hand-written queries. **The outcome:** Claude Code lists catalog assets, follows lineage and returns read-only SQL results, limited to the permissions of the token's creator. ### Steps 1. Start the OLCF platform from the quickstart. 2. Create an MCP token for your own user in the OLCF console. 3. Add the OLCF MCP endpoint to Claude Code using that token. 4. Ask Claude Code to list the sample tables and run a read-only SQL count on one of them. ### Prompt to paste into your agent ```text Use the open-lakehouse-compute-fabric repo (jddevtime/open-lakehouse-compute-fabric) through its MCP endpoint. List the catalog assets in the sample Retail lakehouse, then run one read-only SQL count on
and return the result. Verify that the query was rejected if you ask for any write statement. ``` ### Caveat The README names Claude Desktop and Claude Code as MCP clients, but the exact MCP setup steps are not in the excerpt, so confirm the endpoint settings in the docs before connecting. Risk, credentials: The MCP token carries the permissions of the person who created it, so store it securely and revoke it when the agent no longer needs access. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/jddevtime-open-lakehouse-compute-fabric/let-claude-code-query-the-catalog/ ## How to run a Fabric notebook on your own Spark hardware with Open Lakehouse Compute Fabric Repo: [jddevtime/open-lakehouse-compute-fabric](https://github.com/jddevtime/open-lakehouse-compute-fabric) | Difficulty: Advanced | Time: 240 minutes | Basis: From the README Send a Fabric notebook to your own Spark workers with olcf.run_remote() and read OneLake directly. **Who it is for:** A Fabric data engineer who wants heavy or sensitive notebook compute to run inside their own network instead of consuming Fabric capacity. **The problem:** Large or sensitive Spark jobs in Fabric use Fabric capacity, and some workloads should stay on hardware the team controls. **The outcome:** A Fabric notebook cell that runs on your Spark workers, with status, logs and the exit value returned to the cell and to any calling Fabric pipeline. ### Steps 1. Set up a Fabric tenant you control, following the Fabric integration guide. 2. Attach the olcf_fabric wheel to a Fabric Environment and attach that environment to the notebook. 3. Enroll at least one worker with `olcf worker enroll-token`, then start it on the machine with `olcf-worker run`. 4. Call `olcf.run_remote()` in a notebook cell and review the status and logs it returns. 5. Check the OneLake output table to confirm the write happened. ### Prompt to paste into your agent ```text Use the open-lakehouse-compute-fabric repo (jddevtime/open-lakehouse-compute-fabric) and its Fabric integration guide. Attach the olcf_fabric wheel to a Fabric Environment, call olcf.run_remote() on , and report the returned status and exit value. Verify that the target OneLake table has the expected row count after the run. ``` ### Caveat The README says the project is not Microsoft's runtime, claims no parity with Fabric's proprietary optimisations, and is not affiliated with Microsoft, so expect some notebooks to be rejected by the compatibility analyzer. Risk, credentials: Fabric and Entra setup and worker enrollment tokens grant access to your OneLake data, so scope them tightly and keep them out of notebooks. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/jddevtime-open-lakehouse-compute-fabric/run-fabric-notebook-on-own-spark/ ## How to build a Spark cluster from your own machines with Open Lakehouse Compute Fabric Repo: [jddevtime/open-lakehouse-compute-fabric](https://github.com/jddevtime/open-lakehouse-compute-fabric) | Difficulty: Advanced | Time: 180 minutes | Basis: From the README Enroll several machines as workers and run Spark jobs on the combined Standalone cluster. **Who it is for:** A platform engineer with spare servers, VMs or a rack who needs Spark capacity without moving data to a managed service. **The problem:** Spark jobs need more compute than one machine provides, and the team wants to use hardware it already owns. **The outcome:** A Spark Standalone cluster built from enrolled machines, with a Spark runtime installed and jobs placed only on nodes where that runtime is ready. ### Steps 1. Install a Spark runtime with `olcf runtime install --with-jdk`. 2. Generate an enrollment token with `olcf worker enroll-token`. 3. On each machine that should join, run `olcf-worker run` with the token. 4. Submit a test job to the cluster from the console or CLI and check that it runs on more than one worker. ### Prompt to paste into your agent ```text Use the open-lakehouse-compute-fabric repo (jddevtime/open-lakehouse-compute-fabric). Install the Spark runtime with `olcf runtime install --with-jdk`, enroll workers, and report the cluster members. Verify that a test job runs and that the worker list shows each machine as enrolled. ``` ### Caveat The bundled development signing key is public, so re-sign the runtime with your own key before trusting a production installation. Risk, credentials: Enrollment tokens let a machine join the cluster, so generate them one at a time and revoke any that leak. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/jddevtime-open-lakehouse-compute-fabric/build-spark-fabric-across-own-machines/