The following table shows the format of returned data when retrieving it in a character format:
| Period SQL Type | Conversion to SQL_C_CHAR or SQL_C_WCHAR |
|---|---|
| SQL_PERIOD_DATE | The result is a character string: (‘yyyy-mm-dd’, ‘yyyy-mm-dd’) where the beginning and ending date bounds are converted to the ‘yyyy-mm-dd’ format that is always used by the ODBC specification when converting date SQL data (SQL_TYPE_DATE) to the character C data. Example: (‘2007-05-01’, ‘2007-06-01’) |
| SQL_PERIOD_TIME | The result is a character string in the following format: (‘hh:mi:ss[.f....]’, ‘hh:mi:ss[.f...]’) where the beginning and ending time bounds are converted to the “hh:mi:ss” format that is always used by the ODBC specification when converting time SQL data (SQL_TYPE_TIME) to character C data. The “[.f...]” indicates the optional fractional seconds. It is only included if the Period timestamp precision is non-zero. Up to 6 digits can be used for fractional seconds (ODBC allows for 9, but Period timestamps are limited to 6). Example: (‘06:20:21’, ‘17:12:00’) |
| SQL_PERIOD_TIME_WITH_TIME_ZONE | The result is a character string in the following format: (‘hh:mi:ss[.f...]+-hh:mi’’, ‘hh:mi:ss[.f...]+-hh:mi’’) where the beginning and ending time bounds are converted to the “hh:mi:ss+-hh:mi’” format that is used by ODBC Driver for Teradata when converting TIME WITH TIME ZONE to character C data. Example: (‘06:20:21.22+02:00’, ‘17:12:00.22+02:00’) |
| SQL_PERIOD_TIMESTAMP | The result is a character string: (‘yyyy-mm-dd hh:mi:ss[.f...]’, ‘yyyy-mm-dd hh:mi:ss[.f...]’) where the beginning and ending timestamp bounds are converted to the “yyyy-mm-dd hh:mi:ss[.f....]” format that is always used by the ODBC specification when converting timestamp SQL data to character C data. Example: (‘2007-07-04 22:11:43.37’, ‘2007-07-04 22:11:50.00’) |
| SQL_PERIOD_TIMESTAMP_WITH_TIME_ZONE | The result is the same as SQL_PERIOD_TIMESTAMP with the addition of the ‘+-hh:mi’ as described for SQL_PERIOD_TIME_WITH_TIME_ZONE. |
ODBC Driver for Teradata returns SQL_ERROR from the conversion if the receiving buffer is too small. This is unlike DateTime types where a conversion to C character can result in SQL_SUCCESS_WITH_INFO and a truncated string. The reason for the difference is that a truncated DateTime string might be useful for the applications (for example: a truncated SQL_DATE might be used to obtain just the year), but a truncated Period string is not.