Re: Version 6 - Floating Point Bug?

In comp.databases.oracle I previously wrote:
:
: I am attempting to insert a floating point number (zero and non-zero)
: into a table from a Pro*C program, and am getting a core dump,
: in the Oracle generated code. We feel we've checked everything,
: declarations, tables, etc.
:
: Here's my environment:
: Sun OS 4.1.1, sun model 3
: ORACLE RDBMS V6.0.30.3.1
: PL/SQL V1.0.31.0.2 - Production
: ORACLE Precompiler: Version 1.3.15.1.2
:
:
: The value to be inserted is in a host array, but we tried
: switching to non-arrays, to no avail.
:
: I am able to insert manually from the SQL> prompt.
:
: I should point out the exact same code works under Oracle 7
: on a Sparc. Unfortunately, our environment requires us to maintain
: different environments for the time being.
:
: Has anyone else out there encountered this type of thing under Version 6?
: Please email me your responses.
:
: Thanks.

A few people responded to the problem. Others have been bitten by the same bug. It was suggested that the floating point data be converted to string format with sprintf(), and then insert that data into the database. Upon retrieval, convert the data back to float. We've done that, and as expected, it works.

------------------------------+------------------------------------
Jim Yorton                    | Voice:      +1 708-632-6695
Motorola Inc.                 |
General Systems Sector        | Internet:   yorton_at_cig.mot.com
Cellular Infrastructure Group | UUCP:       ...!uunet!motcid!yorton