All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

• Upgraded to oracledb 2.0 driver which no longer supports encoding of the connection() function.
• Fixed psycopg binary dependency and install documentation

• New option --log to log erroneous lines in a *.bad file of the same name (ER #46)
• New option --ignore to ignore errors in file and continue loading (ER #46)
• New option --truncate to empty (truncate) the target table before the load (ER #45)
• New option --encoding to specify the file encoding, if not compatible with UTF-8 (bug #50)
• New option --case-insensitive-identifiers to uppercase all identifiers (ER #60)
• New option --quote-identifiers to quote all table and column identifiers (ER #61)
• Release csv2db as a Python package (ER #53)

• Read all input files with UTF-8 encoding by default
• Upgraded to oracledb driver for Oracle Database
• Upgraded to psycopg (3) driver for Postgres
• Remove usage of psycopg2.extras.execute_batch() as psycopg 3 now supports executemany() natively
• Change testing strategy to separate DB loading tests
• Implemented GitHub Action testing

• Raise an error when empty column names in the header are encountered (bug #56)

• Added loading tests for all databases
• Enhanced error reporting (ER #37)

• Added PROTOCOL=TCPIP and AUTHENTICATION=SERVER to the Db2 connect string.
• Using psycopg2.extras.execute_batch() to speed up data load with Postgres (ER #39)

• Convert row to tuple on array append (bug #35)
• Use UTF-8 connection encoding for Oracle (bug #36)

• Interactive password entry (PR #33)

• Change csv2db description from "the CSV command line loader" to "the CSV to database command line loader"

• Introduce Windows .bat file wrapper

• Don't print color codes on Windows (bug #29)

• Introduce csv2db icon
• Check for mandatory command line arguments (ER #14)
• Exit program immediately if no files were found (ER #17)
• More debug statements for file reading operations (ER #15)
• Color coded output (can be turned off via $NO_COLOR) (ER #20)
• Provide failing data record as debug output (ER #19)
• More precise status output (ER #24)
• Change default data type to VARCHAR(1000) (ER #26)
• Return code to indicate data loading issue (ER #25)
• Introduce SQL Server support (ER #11)

• Pass correct exit return codes (bug #18)
• Autocommit is explicitly turned off (bug #22)
• Record list cleared on error (bug #23)
• Preserve correct order of column list (bug #21)

• Put /*+ APPEND_VALUES */ hint at right position immediately after the INSERT keyword (bug #12)

• Provide support for direct-path loading with Oracle DB (ER #8)
• Provide support for custom quote characters (ER #10)

• Support for user-defined column separator

• Don't execute batch flush on empty data set (bug #6)

• Get correct source path for shell wrapper (bug #4)

• Db2 LUW support
• Version numbering in program help

• Set default port for all database types (Oracle: 1521, MySQL: 3306, PostgreSQL: 5432, Db2: 50000).