Changes in version 2018-10-31                      

  - ark()'s default keep-open method would cut off header names for
    Postgres connections (due to variation in the behavior of SQL
    queries with LIMIT 0.) The issue is now resolved by accessing the
    header in a more robust, general way.

                     Changes in version 2018-09-27                      

  - unark() will strip out non-compliant characters in table names by
    default.
  - unark() gains the optional argument tablenames, allowing the user to
    specify the corresponding table names manually, rather than
    enforcing they correspond with the incoming file names. #18
  - unark() gains the argument encoding, allowing users to directly set
    the encoding of incoming files. Previously this could only be set by
    setting options(encoding), which will still work as well. See FAO.R
    example in examples for an illustration.
  - unark() will now attempt to guess which streaming parser to use (e.g
    csv or tsv) based on the file extension pattern, rather than
    defaulting to a tsv parser. (ark() still defaults to exporting in
    the more portable tsv format).

                     Changes in version 2018-09-11                      

  - Remove dependency on utils::askYesNo for backward compatibility, #17

                     Changes in version 2018-08-20                      

  - Ensure the suggested dependency MonetDBLite is available before
    running unit test using it.

  - Overwrite existing tables of same name (with warning and interactive
    proceed) in both DB and text-files to avoid appending.

                     Changes in version 2018-08-11                      

  - Added a NEWS.md file to track changes to the package.
  - Log messages improved as suggested by @richfitz
  - Improved mechanism for windowing in most DBs, from @krlmlr #8
  - Support pluggable I/O, based on @richfitz suggestions #3, #10

                 Changes in version 0.0.19 (2026-02-09)                 

  - patch tsv read

                 Changes in version 0.0.18 (2024-01-15)                 

  - patch test infrastructure for handling soft dependency on arrow

                 Changes in version 0.0.17 (2024-01-08)                 

  - patch test infrastructure for Windows

                 Changes in version 0.0.16 (2022-09-09)                 

  - patch for local_db() by defaulting path to subdir.
  - update roxygen

                 Changes in version 0.0.15 (2022-02-15)                 

  - Added window-parallel option for ark'ing large tables in parallel
  - More conditional testing on M1/arm Mac

                 Changes in version 0.0.14 (2021-10-18)                 

  - Patch for test suite for Solaris. arrow package installs on Solaris,
    but functions do not actually run correctly since the C++ libraries
    have not been set up properly on Solaris.

                 Changes in version 0.0.13 (2021-10-15)                 

  - Added ability to name output files directly.
  - Add warning when users specify compression for parquet files.
  - Added callback functionality to the ark function. Allowing users to
    perform transformations or recodes before chunked data.frames are
    saved to disk.
  - Added ability to filter databases by allowing users to specify a
    "WHERE" clause.
  - Added parquet as an streamable_table format, allowing users to ark
    to parquet instead of a text format.

                 Changes in version 0.0.12 (2021-04-05)                 

  - Bugfix for arkdb

                 Changes in version 0.0.11 (2021-03-13)                 

  - Make cached connection opt-out instead of applying only to
    read_only. This allows cache to work on read-write connections by
    default. This also avoids the condition of a connection being
    garbage-collected when functions call local_db internally.

                 Changes in version 0.0.10 (2021-03-05)                 

  - Better handling of read_only vs read_write connections. Only caches
    read_only connections.
  - Includes optional support for MonetDBLite

                 Changes in version 0.0.8 (2020-11-04)                  

  - Bugfix for dplyr 2.0.0 release

                 Changes in version 0.0.7 (2020-10-15)                  

  - Bugfix for upcoming dplyr 2.0.0 release

                 Changes in version 0.0.6 (2020-09-18)                  

  - Support vroom as an opt-in streamable table
  - Export process_chunks
  - Add mechanism to attempt a bulk importer, when available (#27)
  - Bugfix for case when text contains # characters in base parser (#28)
  - Lighten core dependencies. Fully recursive dependencies include
    only 4 non-base packages now, as progress is now optional.
  - Use "magic numbers" instead of extensions to guess compression type.
    (NOTE: requires that file is local and not a URL)
  - Now that duckdb is on CRAN and MonetDBLite isn't, drop built-in
    support for MonetDBLite in favor of duckdb alone.