System.OverflowException in OracleDataReader.GetDecimal

Question

The following code
Dim conn As OracleConnection = ...
Dim ssql As String = "select to_date('2003-04-21 09.54','YYYY-MM-DD HH.SS') - to_date('2003-04-21 09.57','YYYY-MM-DD HH.SS') foo from dual"
Dim cmd As New OracleCommand(ssql, conn)
Dim dr As OracleDataReader = cmd.ExecuteReader
dr.read
Dim val As Object = dr(0)
results in the following exception
System.OverflowException: Arithmetic operation resulted in an overflow.
at Oracle.DataAccess.Types.DecimalConv.GetDecimal(IntPtr numCtx)
at Oracle.DataAccess.Client.OracleDataReader.GetDecimal(Int32 i)
at Oracle.DataAccess.Client.OracleDataReader.GetValue(Int32 i)
at Oracle.DataAccess.Client.OracleDataReader.get_Item
David

Answer 1

Of course!
Thanks,
But that means that this query will behave the same way:
"select 1/3 from dual".
And so almost any calculation executed by the ORACLE could sporatically generate an OverflowException.
This seems to contradict the behavior indicated in the ODP.NET docs:
"Potential Data Loss
The following sections provide more detail about the types and circumstances
where data can be lost.
Oracle NUMBER Type to .NET Decimal Type
The Oracle datatype NUMBER can hold up to 38 precisions whereas .NET Decimal
type can hold up to 28 precisions. If a NUMBER datatype that has more than 28
precisions is retrieved into .NET decimal type, it loses precision."
It says that the data will loose precision, not throw an OverflowException.
The real kicker (and the way I found this) is not in directly accessing the datareader, but rather in the OracleDataAdapter. OracleDataAdapter.fill, at least, should truncate the NUMBER without complaint.
David