Archives of the TeradataForum
Message Posted: Tue, 02 Aug 2005 @ 15:03:53 GMT
| Subj: | | Re: Writing UDF for Numeric Datatypes |
| |
| From: | | Dieter Noeth |
Alexander T wrote:
| | Does anyone know the internal storage for Numeric Datatypes. | |
SQL Reference
Data Types and Literals
Chapter 3: Decimal/Numeric Data Types
DECIMAL/NUMERIC Data Types
"Decimal numbers are scaled by the power of ten equal to the number of fractional digits. The number is stored as a two�s complement
binary number in one, two, four, or eight bytes. The number of bytes used for a decimal value depends on the total number of digits in that
value."
| | I want to write a function which takes two numeric datatypes and returns the greater one or 0 according to a condition. | |
| | It should be able to support any numeric with variable precisions. | |
| | How will I will be able to do that ? | |
AFAIK you'll have to write overloaded UDFs for different datatypes.
| | Also I would like to write a to_char function which takes a numeric type and returns the result in varchar. | |
| | The cast function will return with a dot. I don't need that . | |
| | It should work exactly as the to_char function in oracle. | |
| | If we have column with type NUMERIC (10, 5) and the actual value in the column is say 105 | |
| | Then in Teradata will return in teradata as 105.00000 but i need the value as 105 itself. | |
You can do a lot of to_char variations with a format:
select trim(105.00000 (format 'Z(5)'))
Dieter
| 
