NEWS

---

dataset 0.4.1 (2025-11-16)

This release strengthens the handling of semantically enriched vectors and improves coercion across base R and tidyverse workflows.

Enhancements

• New S3 methods for semantically enriched logical, Date, and POSIXct types.
• Expanded coercion support:
  • as_numeric(), as_character(), as_logical(), as_factor()
  • Optional preservation of semantic metadata (label, unit, concept, namespace).
• Rewritten coercion logic for all defined() vector types, ensuring stable and predictable behaviour.

dataset_df improvements

• as.data.frame.dataset_df() and as_tibble.dataset_df():
• Correct handling of numeric, character, factor, and date-time columns.
• Label-aware coercion for categorical variables.
• Clear separation between attribute stripping and preservation.

Testing and robustness

• Significant increase in test coverage, including tests for all coercion paths, metadata stripping, and temporal types.
• Improved error messaging for invalid type coercion.
• More consistent printing and formatting of defined vectors.

This update improves reliability, consistency, and interoperability of semantically enriched datasets in R.

dataset 0.4.0 (2025-08-26)

A new CRAN release with much improved unit testing and documentation to meet the rOpenSci standards and better methods for the main s3 classes of the package.

• Rewritten vignettes.
• Improved print, summary methods for dataset_df and defined.
• Better handling of multible contributors in bibrecord.
• A new dataset_to_triples and xsd_convert for better serialisation.
• A better handling of empty nodes in RDF.
• Many bug fixes in the way semantic information is translated to RDF.
• var_labels() now similar to labelled::var_lables() behavior, generally haven_labelled_defined as an s3 class works better in the tidyverse.
• New bibliographic helper functions for dataset_format() and contributor().
• Countless small bug fixes to convert to various metadata schemas edge cases, like missing contributors, formatted subjects, etc.
• Better handling of structured metadata with subject()

dataset 0.3.4027

• The new bibrecord() class is handles is the superclass of the dublincore and datacite() classes; these classes have a new print method and they are conforming the current library standard DCTERMS and current repository standard DataCite; unlike utils::bibentry(), they handle contributors and their roles, identifiers, and many other attributes.
• Breaking change: the definition metadata field in the defined() class is changed to the more understandable concept name.
• The defined() vectors print nicely, and the dataset_df() class is more readable, too.
• The missing examples are present, including examples on the use of the semantically richer orange_df example dataset.
• Many code quality improvements and new tests.

dataset 0.3.4023

• Changed iris_df to orange_df in all examples.
• xsd_convert() handles difftime classes and edge cases.
• Small errors fixed in examples.
• Test coverage increased.
• The master branch is renamed to main.

dataset 0.3.4021

• Added support for generic vector methods: length(), head(), tail(), as.vector(), as.list(), and subsetting ([, [[).
• Implemented comparison methods (==, <, >, etc.) that operate on the underlying data while maintaining semantic integrity.
• Introduced custom print() and format() methods that summarise metadata (label, unit, definition) in a concise and human-readable manner.
• Improved the summary() method for defined vectors to display variable metadata and integrate seamlessly with base R statistics.
• Enhanced the c() method to validate compatibility across all semantic attributes (label, unit, definition, namespace) before concatenation.
• Extended vignette with richer examples and explanations of semantic validation, namespaces, and metadata access.
• compare_creators() internal function to add all creators to joined datasets.

This update significantly improves the usability and robustness of semantically enriched vectors in both interactive and programmatic workflows.

dataset 0.3.9 (2025-05-25)

• New CRAN release with many bug fixes, and improvements from peer-review.
• The definition attributes is renamed to concept.
• Improved printing for defined and dataset_df classes.
• Improved compatibility and coercion methods for base R character and numeric types.
• A clearer bibrecord class for extending utils::person and utils::bibentry classes for more modern and cleaner bibliographic references.

dataset 0.3.4 (2024-12-23)

• New release on CRAN.

dataset 0.3.0 (2024-01-08)

• Released on CRAN.
• 0.3.1. Is a minor bug fix with units test on old R releases. It does not affect the functionality of the package.

dataset 0.2.9

• dataset_ttl_write(): write datasets to turtle format;
• with helper functions get_prefix(), get_resource_identifier(), xsd_convert(), and dataset_to_triples().

dataset 0.2.8

New vignettes on

• [x] Richer Semantics for the Dataset's Variables

dataset 0.2.7 (2023-12-08)

• Released on CRAN

The devel branch contains new code that is not is validated, but as a whole the package is not working consistently.

dataset 0.2.6

• All tests are passing, all examples are running.

dataset 0.2.5

• datacite() has a new interface and an as_datacite() retrieval version. See the Working with DataCite Metadata vignette.
• dublincore() has a new interface and an as_dublincore() version. See the Working with Dublin Core Metadata vignette.

dataset 0.2.4

All tests are passing but documentation is not rewritten yet.

dataset 0.2.3

new subject class for recording subjects

dataset 0.2.2

New s3 classes for DataCite and Dublin Core bibliographic entries.

dataset 0.2.1 (2023-03-18)

A minor correction to avoid vignettes downloading data from the Eurostat data warehouse on CRAN. Small readability improvements in the vignette articles.

dataset 0.2.0 (2022-12-14)

• New methods for the dataset() s3 class: print.dataset(), summary.dataset(), subset.dataset, [.dataset, as.data.frame().
• New vignette on how to use the dataspice package programmatically for publishing dataset documentation.
• Released on CRAN.

dataset 0.1.9 (2022-12-02)

• Incorporating minor changes from the rOpenSci and CRAN peer-reviews.

dataset 0.1.7

• After reviewing CRAN submission comments, and correcting documentation issues, submitted to rOpenSci for review before re-submitting to CRAN.

dataset 0.1.6.0001

• Add dataset_local_id() and dataset_uri() to the dataset functions.

dataset 0.1.6

• A release candidate on CRAN after small documentation improvements.

dataset 0.1.4

Development version available on Zenodo.

• dataset_export() is implemented with filetype = 'csv'.
• Replacement functions are added to simple properties identifier(), publisher(), publication_year(), language(), description(), datasource_get() and datasource_set() [to avoid confusion with the base R source() function], geolocation(), rights(), version().
• Functions to work with structured referential metadata: dataset_title(), subject(), subject_create().

dataset 0.1.3

• Vignette articles started to develop and consult the development plan of the project. See From dataset To RDF, Export and Publish A dataset Object, Datasets with FAIR metadata.
• New functions: download_dataset(), datacite(), and the dataset() constructor.

dataset 0.1.2

• The definition of the dataset() class, an improved data.frame (tibble, DT) R object with standardized structure and metadata.
• Adding and reading DublinCore metadata and DataCite mandatory and recommended FAIR metadata metadata.

dataset 0.1.0

First development version release.

• Added the Motivation of the dataset package vignette article, which is later replaced with Design Principles & Future Work Semantically Enriched, Standards-Aligned Datasets in R.