DataCharter data checks — GitHub Action

Run contract-governed data quality checks in CI. This action runs DataCharter's test (data assertions declared in your charter.yaml) and drift (schema + PII drift against a committed baseline), and fails the job when they fail — so a broken contract blocks the PR, not the pipeline.

Results are written to the job's step summary as a ✓/✗ report.

Usage

name: data-checks
on: [push, pull_request]

jobs:
  data:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v5
      - uses: datacharter/datacharter-action@v1
        with:
          workspace: .           # directory containing charter.yaml
          run-drift: "true"      # also check schema/PII drift

Declare assertions in your charter.yaml:

tests:
  id_not_null:  { type: not_null, relation: orders, column: id }
  id_unique:    { type: unique, relation: orders, columns: [id] }
  region_valid: { type: accepted_values, relation: orders, column: region, values: [US, EU] }
  has_rows:     { type: row_count, relation: orders, min: 1 }
  totals_ok:    { type: expression, relation: orders, expression: "amount >= 0" }

For the drift check, commit a baseline once (datacharter drift . --update writes .datacharter/schema.json); CI then fails when tables, columns, or PII declarations drift from it.

Inputs

Input	Default	Description
workspace	.	Path to the workspace (the directory containing charter.yaml).
version	latest	Pin a datacharter version, e.g. 0.10.4.
run-tests	true	Run the tests: assertions.
run-drift	false	Run the schema/PII drift check.
select	—	Run a single test by name.

Databases and credentials

File-based sources (CSV, Parquet, Excel, SQLite, DuckDB) committed to the repo work with no setup. For live databases, charter.yaml credential references (${VAR}) resolve from the environment — pass them as workflow secrets:

      - uses: datacharter/datacharter-action@v1
        env:
          PG_PASSWORD: ${{ secrets.PG_PASSWORD }}

Every check runs through DataCharter's read-only engine: CI can never write to your sources.

Links

• Main project: https://github.com/datacharter/datacharter
• Docs: https://datacharter.dev · CLI reference
• Blog: https://datacharter.dev/blog/

Apache-2.0.