Verify a Data Standard
Run a dataset's Data Standard checks on demand with the Soda CLI.
Verifying a Data Standard runs its checks against real data and reports whether the data meets the rule. Active standards are verified automatically on their schedule in Soda Cloud. You can also trigger a verification on demand, either from the dataset's page in Soda Cloud or from the Soda CLI (useful for CI/CD pipelines, custom orchestration, or reproducing a scheduled run locally).
Verify on demand in Soda Cloud
You run a standard's checks from the dataset's page, not from the standard itself. Use the Verify button on the dataset to run its checks on demand.
By default, Verify runs both the dataset's Data Standard checks and its Data Contract checks in one go. To run just one, use the dropdown next to the button and choose the Data Standard checks (or the contract checks) individually.
Verify a dataset's standards from the CLI
Use soda data-standard verify with --dataset (-d) to fetch and run all active standards that apply to a specific dataset, as resolved by their scopes in Soda Cloud.
soda data-standard verify \
--dataset datasource/db/schema/customers \
--data-source ds_config.yml \
--soda-cloud sc_config.ymlThis command:
Fetches the active standards for the dataset from Soda Cloud, already resolved with the dataset and its active columns
Connects to your data source using the configuration file
Runs the checks and returns a pass/fail result
Because the standards are resolved by Soda Cloud, this run mirrors exactly what the scheduled cloud-managed verification does.
Publish results to Soda Cloud
Add --publish (-p) to send verification results to Soda Cloud for visibility, history, and alerting.
Learn how to connect the CLI to Soda Cloud: Connect to Soda Cloud.
Override variables
If a standard declares variables, pass values at verification time with --set:
Run a subset of checks
Use --check-paths (-cp) to run only specific checks, identified by their YAML path. This is useful for targeted debugging of a single check.
A dataset can have several standards active at the same time, so a check path names which standard the check belongs to as well as where it sits inside that standard:
Pass several paths to run a subset that spans standards:
Because . and : separate the parts of a check path, a Data Standard's name cannot contain either character.
Command options
--dataset, -d
A Soda Cloud dataset identifier. Fetches and verifies that dataset's active standards. Required.
--data-source, -ds
One or more data source configuration file paths. The first is treated as primary. Required.
--soda-cloud, -sc
Path to a Soda Cloud configuration file. Required.
--publish, -p
Send the verification results to Soda Cloud.
--set
Set a variable value, in the form --set NAME=VALUE. Repeatable.
--check-paths, -cp
One or more check paths to run a subset of the checks. Each path names its standard: data-standard.<name>:<path>.
--diagnostics-warehouse, -dw
Path to a diagnostics warehouse configuration file.
--verbose, -v
Display detailed logs during execution.
Next steps
Review verification results across datasets: Review results
Last updated
Was this helpful?
