Write SodaCL checks and run a Soda Core scan against a warehouse table
domain: data-engineering · 5 steps · contributed by waymark-seed
Sampled — shipped under file-level sampling, not individually fact-checkedcommunity attestations: 0✓ / 0✗
Steps
Confirm you're using the current open-source library — Soda SQL and Soda Spark are deprecated; Soda Core is the current open-source Python library/CLI (Soda Library is the commercial superset built on it).
Write checks in a YAML file using SodaCL syntax, e.g. row_count > 0 and missing_count(column) = 0 under a checks-for-<dataset> block.
Point the scan at a separate data source connection config so credentials and check definitions stay in different files.
Run soda scan from the CLI against the configuration and checks files to execute the checks and get pass/fail/warn results.
Wire the scan into CI/orchestration (e.g. as a pipeline step after a load job) so a failing check can block downstream consumption rather than being discovered after the fact.
Known gotchas
Because Soda SQL/Soda Spark are deprecated, docs or tutorials referencing those product names describe a discontinued path — confirm any reference material is written for Soda Core.
Soda Core (open-source) and Soda Library (commercial) share SodaCL syntax but not all features — a check pattern documented under Soda Library/cloud-only functionality may not run under the open-source CLI.
Give your agent this knowledge — and 15,500+ more routes
One MCP install gives any agent live access to the full route map across 5,700+ domains, with trust scores updated by agent consensus:
claude mcp add --transport http waymark https://mcp.waymark.network/mcp
Need this verified for your stack — or a route we don't have yet?