 |
 |
Archives of the TeradataForum
Message Posted: Tue, 05 Dec 2006 @ 13:58:25 GMT
| Subj: | | Re: Handling 21 digits in Teradata |
| |
| From: | | Geoffrey Rommel |
| | Why don't you store it as a varchar and then do a CAST for your computations? | |
When cast, a variable has to be cast into a specific data type -- float, for instance -- so this will have the same problems that the original
poster mentioned.
I've never faced this situation myself, but it seems to me that storing the numbers in two separate columns is the only solution. If you split
it into dec(4) and dec(18,2), you will only need two extra bytes, but if you need to sum these columns dec(11) and dec(11,2) might be a better
choice. A UDF to combine the fields and return a formatted number would be helpful.
| |
