NULL Byte Support in PostgreSQL Character Columns

With z/OS DB2, you can store the NULL byte (x'00') into character columns; however, PostgreSQL does not allow this. To work around this limitation:

• Store your tables in a UTF-8 database

  In a UTF-8 database, the NULL byte is replaced with UTF-8 value (x'C680') when doing DML in an SQL program, when using database migration tools provided by AWS, and when using the MBDT LOAD and UNLOAD utilities.

• Compile programs with the SQL(ALLOWNULLCHAR) compiler directive
  Important: If your application requires NULL byte support for a CHAR field, additional configuration is required. See ALLOWNULLCHAR for details.
• Set environment variable PGCLIENTENCODING=UTF8
• Use the PostgreSQL ODBC Driver (UNICODE)
• Set your Windows or Linux system locale to appropriate the Single-byte Character Set (SBCS) value:
  • For Windows, the system locale is defined in the Windows Control Panel. For example, English(United States) uses Windows SBCS code page 1252.
  • For Linux, the system locale is defined by the value of the LANG environment variable. For example, setting LANG=en-US.iso8859-1 or LANG=en_US.iso8859-15 uses Linux SBCS code page 819 (non EURO) or 923 (EURO).
• For both Windows and Linux platforms, we provide special conversion tables containing SBCS code pages that map NULL byte to UTF-8 value (x'C680'). You must install these conversion tables and configure your system to find them. For complete instructions, see To install CCSID tables for HCOPG.