Skip to contents

Changelog

Source: NEWS.md

nanoarrow 0.7.0

CRAN release: 2025-07-03

  • Add zstd decompression support to R package (#733)
  • Support native creation of more numeric Arrow arrays from integer vectors (#697)
  • Support matrix objects as fixed-size-list arrays (#692)
  • Ensure that python is used on Windows when running bootstrap.R (#792)
  • Update vctrs extension name to reflect implementation change (#752)
  • sub-day precision Date should be floored when treated as integer (#674)

nanoarrow 0.6.0

CRAN release: 2024-10-13

  • Add float16 support for R bindings (#650)
  • Implement string view support in R bindings (#636)
  • Allow opt-out of warning for unregistered extension types (#632)
  • Add bindings to IPC writer (#608)
  • Avoid flatcc aligned_alloc() call when compiling R package (#494)
  • Use JSON in experimental R vctrs extension type (#533)

nanoarrow 0.5.0

CRAN release: 2024-05-26

  • Add experimental nanoarrow_vctr to wrap a list of arrays (#461)
  • Add bindings for IPC reader (#390)
  • Fix tests for platforms where arrow dataset and/or zip is not available (#415)
  • Remove unnecessary package name call (#381)

nanoarrow 0.4.0

CRAN release: 2024-02-01

  • Fix source links from pkgdown site (#315).
  • Provide LinkingTo headers for extension packages (#332).
  • Add more nanoarrow_array_stream generics (#349).
  • Add conversion from integer type to character() (#345).
  • Ensure simple list()s can be converted without arrow installed (#344).

nanoarrow 0.3.0.1

CRAN release: 2023-12-08

  • Ensure wrapper array stream eagerly releases the wrapped array stream (#333).

nanoarrow 0.3.0

CRAN release: 2023-09-29

  • Use classed warnings to signal that a lossy conversion occurred (#298)
  • Add support for bit64::integer64() conversions (#293)
  • Implement extension type registration/conversion (#288)
  • Implement dictionary conversion (#285)
  • Ensure ordered is reflected in na_dictionary() (#299)
  • Warn for possibly out of range int64 -> double conversions (#294)
  • Support map conversion to R vector (#282)
  • Don’t link to arrow package R6 class pages (#269)
  • Use basic_array_stream() to improve array stream to data.frame conversion (#279)

nanoarrow 0.2.0-1

  • Don’t link to arrow package R6 class pages (#269)

nanoarrow 0.2.0

New features

  • Improve printing and conversion of buffers (#208)
  • Add enum ArrowType buffer_data_type member to struct ArrowLayout (#207)
  • Implement ListChildOffset function (#197)
  • Add ability to deterministically run a finalizer on an array stream (#196)
  • Union array support (#195)
  • Add ArrowArrayStream implementation to support keeping a dependent object in scope (#194)
  • Add as_nanoarrow_array() implementation that does not fall back on arrow::as_arrow_array() everywhere (#108)
  • Create nanoarrow_array objects from buffers (#105)
  • Implement infer schema methods (#104)
  • Create and modify nanoarrow_schema objects (#101)

Bugfixes

  • Fix convert_array_stream() for non-record batch stream with zero batches (#212)
  • clear release in EmptyArrayStream::release_wrapper (#204)
  • Release streams when calling as.vector() or as.data.frame() (#202)
  • Don’t invoke undefined behaviour in conversions to/from Arrow (#167)
  • Use strict prototypes in all internal C functions (#151)
  • Don’t memcpy NULL when converting buffer to raw (#149)