Data Standards

Write data quality checks once and deploy them across many datasets at scale with Data Standards.

Data Standards are currently available as a preview feature. Contact [e-mail] for early access.

What is a Data Standard?

A Data Standard is a reusable set of data quality checks that you write once and deploy across many datasets at the same time.

Where a Data Contract describes the expectations for a single dataset, a Data Standard expresses an expectation that should hold everywhere a rule applies, for example: "every dataset that contains an email column must have valid, non-missing email addresses," or "every table in the finance schema must have a passing schema and row_count check."

Data Standards are how teams enforce organization-wide data quality governance without authoring and maintaining a separate contract for every dataset. You define the rule centrally, target the datasets it should apply to, and Soda keeps it enforced as your data landscape grows.

The Data Standards overview, listing each standard with its owner, matched datasets, check results, compliance, and status.

Data Standards reuse the same check engine and YAML check language as Data Contracts.

If you already know how to write contract checks, you already know how to write a Data Standard: the checks (missing, invalid, duplicate, schema, row_count, freshness, and so on) are identical.

Data Standards vs. Data Contracts

Data Standards and Data Contracts are complementary. They share the same check language but solve different problems.

Data Standard	Data Contract
One expectation that should hold in all datasets it applies	A tailored agreement for the expectations in one dataset
Written once, deployed across a fleet	Written per dataset
Scope: many datasets at once, selected by a rule	Scope: a single dataset
When onboarding, matching datasets inherit the standard automatically	Suitable for new datasets that need their own contract
Best for organization-wide governance enforced consistently at scale	Best for a dataset that deserves its own bespoke rules

Aspect	Data Standard	Data Contract
Scope	Many datasets at once, selected by a scope rule	A single, specific dataset
Intent	Enforce a quality rule in all datasets it applies	Define what "good data" means for this dataset
Authored by	Data governance / platform teams setting org-wide policy	The team that owns the dataset
Identity	Named rule that fans out to every matching dataset	Bound to one dataset's fully qualified name
Typical use	"All PII columns must be valid", "every table must pass schema" Matching datasets inherit a standard automatically	Producer/consumer agreement on a key table A new dataset needs its own tailored contract

Use a Data Contract when a dataset deserves its own tailored agreement. Use a Data Standard when you want the same rule applied consistently across a fleet of datasets, whether it is a whole contract, or a check for a specific field.

How to implement a Data Standard

Creating and managing Data Standards requires the Manage data standards permission. Data Standards must also be enabled for your organization.

Learn more about

1

Write the standard

Define the data quality checks you want to enforce, using the same check language as a Data Contract.

See Author a Data Standard.

2

Deploy it at scale

Create the standard in Soda Cloud and give it a scope, a rule that selects which datasets it applies to. Soda links the standard to every matching dataset and re-evaluates the scope as datasets change.

See Deploy at scale.

3

Verify the data

Soda runs the standard's checks against every linked dataset, on a schedule or on demand. You can also trigger a verification for a dataset from the CLI.

See Verify a Data Standard.

4

Review the results

Inspect pass/fail results, history, and diagnostics for each check across every dataset the standard touches.

See Review results.

Lifecycle

A Data Standard moves through a small set of states:

• Draft – created but not yet enforced. No datasets are linked and no scans run.
• Active – enforced. The standard is linked to every dataset that matches its scope, and its checks run on the configured schedule.
• Paused – temporarily suspended. Existing links are kept, but no new scans run until you resume it.

Prerequisites

Before working with Data Standards, make sure:

• You have a Soda Cloud account with the Dataset Standard, Team, or Enterprise license.
• You have connected at least one data source and onboarded datasets. See how to onboard data sources.
• To create or manage standards, your user needs the Manage Data Standards permission.

---

{% if (visitor.claims.plan === 'datasetStandard') %}

You are logged in to Soda and seeing the Dataset Standard license documentation. Learn more about Documentation access & licensing.

{% if (visitor.claims.plan === 'enterprise') %}

You are logged in to Soda and seeing the Team license documentation. Learn more about Documentation access & licensing.

{% if (visitor.claims.plan === 'enterpriseUserBased') %}

You are logged in to Soda and seeing the Enterprise license documentation. Learn more about Documentation access & licensing.

{% if !(visitor.claims.plan === 'enterprise' || visitor.claims.plan === 'enterpriseUserBased' || visitor.claims.plan === 'datasetStandard') %}

You are not logged in to Soda and are viewing the default public documentation. Learn more about Documentation access & licensing.

If you do have a Soda license, make sure to log in to Soda Cloud in this same browser.