Tabular data

A format, Data Package, recognised by intent: a Frictionless Data Package descriptor (datapackage.json, datapackage.yaml), a Table Schema (tableschema.json), or a JSON or YAML file whose $schema or profile names a Data Package profile (datapackage.org, frictionlessdata.io, tabular-data-package) makes the project a tabular data one. A CSV alone is a table of no declared schema and makes nothing (.csv and .tsv are data, never sources). The descriptors stay JSON or YAML, the Data format's, so data-wellformed already covers one that does not parse.

A CI job works with the ecosystem when it runs frictionless, goodtables or csvlint, or the frictionlessdata/frictionless-ci action (once frictionlessdata/repository).

schema-validation

CI validates the tables against their schemas: frictionless validate datapackage.json (a package, a resource, or a table with --schema; describe and extract only read them), or the frictionlessdata/frictionless-ci action, which validates the inquiry, packages, resources or tables its inputs name, and the repository's data packages when told nothing; goodtables validate, its predecessor; csvlint --schema=<table schema> <table> (without a schema, csvlint checks that the CSV is well formed, against nothing).

schema-declared

Every package or resource descriptor declares its profile ($schema, version 2; profile, version 1) and its name. A Table Schema on its own names neither and is not asked.

Not judged

A dvc.yaml alone shows data versioning, not a schema to validate against, and large-files already accepts a file DVC tracks, so it recognises nothing here. The other schema practices of #85 (schema-declared, schema-examples, schema-compatibility) are not yet built. The language practices have no tool for a table and are skipped.