Re: ORACLE 8 NUMERIC PROCESSING BUG?

A copy of this was sent to Peter Hanlon <peter_at_hanlon.co.uk> (if that email address didn't require changing) On Sun, 22 Aug 1999 21:36:53 GMT, you wrote:

>A change has been made to the way that Oracle 8 handles FLOAT columns.
>If you create a table with a FLOAT column and you describe the table,
>the column will be reported back as a NUMBER.
>

this is bug 686159 broken in 8.0, fixed in 8.0.6 and 8.1.6.

>On the face of it this doesn't seem so terrible, except that the values
>that can be stored in a FLOAT are not the same as the values that can
>be stored in a NUMBER. For example a FLOAT(10) can only store 3/4
>digits where as a NUMBER(10) can store 10 digits, so the two data types
>are not equal.
>

true.

[snip]

>
>Another interesting feature is that if you alter a NUMBER(10) column to
>be a FLOAT(10) column it works but doesn’t truncate the values already
>stored in the table or raise an error reporting that the two data types
>are not compatible. If you later attempt to insert a ten digit number
>into the table, it gets truncated. I guess this is sort of correct but
>it makes debugging code a nightmare when your program used to work but
>doesn't anymore.
>

I think that is a bug. It should work just like a varchar2(10) being modified to a varchar2(4) should work. I filed a bug.

[snip]

--
See http://govt.us.oracle.com/~tkyte/ for my columns 'Digging-in to Oracle8i'... Current article is "Part I of V, Autonomous Transactions" updated June 21'st  

Thomas Kyte                   tkyte_at_us.oracle.com
Oracle Service Industries     Reston, VA   USA

Opinions are mine and do not necessarily reflect those of Oracle Corporation Received on Tue Aug 24 1999 - 14:13:32 CDT