NEWS

---

container 1.1.0.9003

• Same as previous version.

container 1.1.0.9002

• Same as previous version.

container 1.1.0.9001

• docs: update interactive usage guide - cleanup old artifacts
• docs: update pkg site; include %in% operator in reference index

container 1.1.0.9000

• Same as previous version.

container 1.1.0 (2025-11-10)

New features

• Extraction ([.Container):

  • Non-standard evaluation (NSE) for ranges and mixed endpoints, e.g. x[a:b], x[1:c], x[d:2], and negative ranges like x[-(a:c)].
  • Multiple indices as separate arguments and list(...) sugar: x[1, "a", 3:5], x[list(1, "a")].
  • Logical indexing improvements with safe recycling and informative warnings when the mask length doesn’t divide the container length; NA treated as FALSE with a warning.
  • .default argument supported to fill unknown/out-of-bounds positive indices and unknown names while preserving order and duplicates.
  • x[[TRUE]] now aligns with base R to return the first element.

• Replacement ([<-.Container):

  • Supports the same extended indexing options as extraction, including NSE ranges, mixed indices, logical masks, and negative (complement) selection.
  • x[] <- v targets all positions; zero-length selections perform no replacement.
  • Adds unknown names when replacing via character indices; numeric indices must be in-bounds.
  • Replacement length is recycled with a warning when not a multiple of the target length.

• Operators & utilities:

  • Generic %in% operator added (#36).
  • Option for shallow copy when converting container to list.

Fixes

• Allow setting names for initially unnamed containers.

Internal / documentation

• Refactored extract indexing into internal helper .get_pos_indices and reused it for replace to ensure consistent semantics.
• Reworked and expanded tests; migrated legacy tests to testthat; improved docs and vignettes (interactive usage and replacement examples).
• Renamed vignettes for CRAN ordering (#33); updated README and vignettes (#30).
• Added News menu (this Changelog) to the package site and links to other packages via R universe.
• Clarified OpsReplace documentation (indexing options, NULL assignment, NSE alphanumeric ranges).

container 1.0.5 (2024-12-02)

• Update pages to bootstrap 5 and github action pipelines (#26)
  • upgrade pages to use bootstrap 5
  • update README
  • remove Date field in DESCRIPTION file
  • fix documentation signaled by CRAN devel checks

container 1.0.4 (2022-12-11)

• fix "don't run" examples
• fix Rd files regarding tidy HTML

container 1.0.1 (2022-02-19)

• prevent duplicated column names when renaming dict.tables
• don't check names in dict.table constructor - allow optional checking

container 1.0.0 (2021-11-06)

Breaking changes

This update brings some big changes and breaks some of the earlier functions. However, with version 1.0.0, from now on, the API will remain stable.

All classes Container, Deque, Set and Dict now have been modified to be more similar in the way elements can be added, extracted, replaced or removed and therefore share most of their operations. The motivation behind this was to provide specialized data structures that enhance base R list but at the same time still 'feel' familiar to R users. As a result, all classes (including Set and Dict) now provide named elements and access via name or integer indices similar to base R list, which should users enable to use any of them right from the start without having to read through the manual.

A new class dict.table is provided, a mixture of data.table and Dict, which extends data.table by Dict operations to enhance data column management.

List of changes

• Set and Dict provide positional access
• for Dict, internally all key-value pairs are stored in a hash-table and the elements are always sorted lexicographically by their keys
• in addition to Set there is a new class OrderedSet, in which all elements are also always sorted.
• objects of any class are now initialized with elements passed via name = value as is done in base R lists.
• add() now accepts multiple elements passed via name = value.
• delete() (formerly remove) and discard() now always work by value. For removing elements by index, now delete_at() and discard_at() are used.
• has() now always works by value. For checking elements by name, now has_name() is used.
• the print() method has been fully revised and now (inspired by the sets package) prints very compact.
• size() is now determined by length()
• count(), pop, and update are now available for all classes
• the S3 method interface now by default provides copy semantics. For reference semantics, functions starting with ref_ are used.

New features

• new classes OrderedSet and dict.table
• all classes support named elements
• names() is used to get and set the names as in base R list.
• extract or replace parts of an object via standard [, [[, and $ operators using character or numeric indices or both at the same time.
• container_options() to customize printing and applied comparison operators
• unpack() to unpack nested container objects
• replace() and replace_at() to replace values and values at indices
• at() and at2() for safe element extraction
• peek_at() and peek_at2() for loose element peeking with default values
• ref_pop() get and remove element at given index
• rename() elements in place
• comparison operators that compare containers lexicographically
• arithmetic operators
• group generic functions for Math and Summary

Deprecated Functions

• empty() - use is_empty() instead
• set() - use setnew() instead
• size() - use length() instead
• sortkey() - keys of Dict objects are now always sorted
• remove() - use delete() instead
• type() - not of use anymore
• values() - use as.list() instead

container 0.3.0 (2018-12-01)

Initial release