Changelog
This document includes all meaningful changes made to the specifications consisting the Data Package Standard. It does not track changes made to other documents like recipes or guides.
v2.0-draft
April 1, 2024
Overview
The Data Package (v2) draft release includes a rich set of the specification improvements accepted by the Data Package Working Group during the active phase of the Data Package (v2) work.
Changes
Specifications
Added source.version property
This change adds a new property to make possible of providing information about source version. Please read more about source.version property.
Made contributor/source.title not required
This change allows omitting title property for the contributor and source objects making it more flexible for data producers.
Data Package
Added contributor.given/familyName
This change adds two new properties to the contributor object: givenName and familyName. Please read more about package.contributors property.
Added contributor.roles property
This change adds a new contributors.roles property that replaces contributor.role. Please read more about package.contributors property.
Fixed version property in Data Package profile
This change adds omitted version property to the Data Package profiles.
Data Resource
Relaxed resource.name rules but keep it required and unique
This change relaxes requirements to resource.name allowing it to be any string. This property still needs to present and be unique among resources. Please read more about resource.name property.
Clarified resource.encoding property
This change updates the resource.encoding property definition to properly support binary file formats like Parquet. Please read more about resource.encoding property.
Forbade hidden folders in paths
This change fixes definition in the Data Resource specification to explicitly forbid hidden folders.
Table Dialect
First version of the specification
This change adds a new specification Table Dialect that superseeds and extends the CSV Dialect specification to work with other formats like JSON or Excel. Please refer to the Table Dialect specification.
Table Schema
Added schema.fieldsMatch property
This change clarifies the default field matching behaviour and adds new modes for matching data source and Table Schema fields. Please read more about schema.fieldsMatch property.
Made any be a default field type
This change makes field type to be any by default and ensures that the field type is not inferred if not provided. Please read more about any type.
Added uniqueKeys property
This change adds uniqueKeys property directly modelled after corresponding SQL feature. Please read more about schema.uniqueKeys property.
Added field.missingValues
This change adds a property that allows to specify missing values individually per field. Please read more about field.missingValues property.
Added list field type
This change adds a new field type list for typed collections, lexically delimiter-based. Please read more about list type.
Added jsonSchema constraint to object and array fields
This change adds a new constraint for the object and array fields. Please read more about constraints.jsonSchema constraint.
Support groupChar for integer field type
This change adds support for providing integers with group chars. Please read more about field.groupChar property.
Extended datetime default format
This change extends default format definition for the datetime field type allowing to provide optional milliseconds and timezone parts.
Supported exclusive constraints
This change adds new exclusiveMinimum and exclusiveMaximum constraints to the Table Schema specification.
Simplified self-referencing in foreign keys
This change allows omitting foreignKey.resource.reference in case of self-referencing. Previously it required setting resource to an empty string.
Discouraged usage of unnecessary union types
This change discourages usage of mixed types for schema.primaryKeys and schema.foreignKeys.fields properties.
Clarified that geopoint is number-based
This changes clarifies that geopoint field type can use floating point numbers for coordinate definitions.
Fixed duration constraint
This change fixes minimum and maximum constraint for the duration field type.
v1.0
September 5, 2017
Please refer to the the Data Package (v1) website.