DB2 Everyplace UNICODE support

On operating systems that support UNICODE (Windows CE, Symbian OS, Windows NT and Windows 2000), DB2 Everyplace takes UNICODE strings only as Input/Output strings. These UNICODE strings are saved as UTF-8 format inside the DB2 Everyplace engine. A UNICODE character might require one to three bytes of storage space after the UTF-8 conversion. A character string stored in a database server such as IBM DB2 Universal Database might require more space when the string is downloaded and stored in DB2 Everyplace on Windows CE devices.

CLI UNICODE interface notes:

The DB2 Everyplace CLI UNICODE functions have a character "W" appended at the end. By defining the macro UNICODE (which is the system default on Windows CE), the regular CLI functions map to the corresponding UNICODE functions automatically. To write portable code, define the macro "UNICODE", and let the system do the conversions.
When UNICODE support is enabled, the data types SQL_C_CHAR, SQL_C_TCHAR, and SQL_C_WCHAR have the same meaning.

Many CLI functions have a string (or buffer) length as an input/output parameter.

For functions with Argument Type as SQLCHAR* (or SQLWCHAR* for the W function), the length is the number of characters. For example:

SQLRETURN   SQLExecDirect    (SQLHSTMT          hstmt,
                              SQLCHAR     FAR   *szSqlStr,
                              SQLINTEGER        cbSqlStr);

UNICODE string L"ABCD" is four characters.

For functions with Argument Type as SQLPOINTER, the length is the number of bytes. For example:

SQLRETURN  SQLGetData       (SQLHSTMT          hstmt,
                             SQLUSMALLINT      icol,
                             SQLSMALLINT       fCType,
                             SQLPOINTER        rgbValue,
                             SQLINTEGER        cbValueMax,
                             SQLINTEGER  FAR   *pcbValue);

The length for the input parameter cbValueMax and output parameter *pcbValue are in bytes. UNICODE string L"ABCD" is eight bytes.

The UNICODE functions can also take SQL_NTS to indicate a NULL-terminated string.

Tips for writing portable code:

Use SQLTCHAR instead of SQLCHAR or SQLWCHAR.
Use the _tcsXXXX functions instead of strXXXX (ANSI) or wcsXXXX (UNICODE). For example, use _tcslen() instead of wcslen() or strlen().
Use _TEXT() ( or TEXT() ) to wrap literal strings. For example, _TEXT("ABCD") can be interpreted as either an ANSI or UNICODE string depending on the macro definition.
Use sizeof(ArrayName)/sizeof(TCHAR) to find out the size of a character array.

For an example, see the Windows CE SampleCLP sample code included with DB2 Everyplace.

Related reference