Release 0.11.1

[eddelbuettel]

This PR updates the package to release 0.11.1. From NEWS.md (with 'lowered' headline levels for display here)

tiledb 0.11.1

• This release of the R package builds against TileDB 2.6.4, but has also been tested against earlier releases, and the development version.

Improvements

• Use of TileDB Embedded was upgraded to release 2.6.4 (Release 0.11.1 #384) following an earlier upgrade to 2.6.2 (Use TileDB 2.6.2 #359)

• Creations of arrays from data.frame objects now supports a mode= argument with values 'ingest', 'schema_only', and 'append' (Add mode=... for data.frame ingestion #360)

• Some unit test and continuous integration code was refactored (Further unit test cleanup #364, Further small CI refinements #375)

• Finalizer use is now simplified taking advantage of an Rcpp change (Simplify finalizer use  #366)

• A new option strings\_as\_factors was added for data.frame retrieval (Support strings_as_factors option #367)

• The arrow C-level interface now uses external pointer objects following Arrow 7.0 (Update use of Arrow C interface  #368)

• Support for memory limits has been extended, and partial reads are using with iterations to complete (Update and extend support for memory limits and iterate on partial reads #371)

• Fragment info reading now account for the __fragments object (Adjust the schema uri retrieval used by fragment info for a new layer #373)

• A nightly test under valgrind has been added; results are reported to slack (Analyse valgrind losses (via awk) and send summary if losses seen #382, Analyse valgrind losses (via awk) and send summary if losses seen (2nd PR) #383)

• UTF-8 string in metadata are now supported (Support UTF-8 strings in metadata values #377)

• Attribute-less arrays can now be created, written, and read (Support attribute-less schemas in creation, write, read #378), also via higher-level accessors (Support array reads returning no attributes #379)

• A plugin for Rcpp has been added (Support Rcpp plugin usage #380)

Bug Fixes

• Array status is now checked before closing (Check array status before array close() call #362)

• Signed and unsigned int64 dimensions are now mapped correctly from 'square-bracket indexing', and the third dimension is recognised (Map (u)int64 indices and restore third dim when indexing #365)

• Domain information could overflow int64_t if an unsigned value was used, this now flips to double (For domain info cast int64 to double if overflowing #370)

• Unit tests for consolidation and vacuuming were update to account for __fragments too (In unit tests change directory for consolidation and vacuuming counts to __fragments #374)

• A unit test was corrected to ensure logical expressions are of length one (Adjust one test to ensure 'length 1' after coercion to logical #381)

Documentation

• A new vignette on data ingestion has been added (Add vignette on data ingestion #357)

• A new vignette on installation options has been added (Add vignette on installation options #358)

• The vignettes are now built using package simplermarkdown (Switch vignettes to simplermarkdown #361)

• Help pages were polished (Add small polish to help pages post recent merges #369)

Deprecations

• The tiledb_dense and tiledb_sparse functions which were deprecated in February 2021 have been removed after a twelve-month grace period.

[shortcut-integration]

This pull request has been linked to Shortcut Story #16193: R package release 0.11.1.

[ihnorton]

🚀 very nice change-list!