4 Replies Latest reply: Jun 7, 2013 2:55 PM by zeroid RSS

    BINARY_DOUBLE and support for double.PositiveInfinity

    zeroid
      I'm using latest ODAC with Entity Framework 4.1. I use the BINARY_DOUBLE type which gets mapped to a .NET double. I use store the results of double arithmetic which mostly works fine, however if the result is Infinitiy (double.PositiveInfinity) then Oracle throws this error when I SaveChanges:

      ORA-01426: numeric overflow

      Since Oracle BINARY_DOUBLE supposedly does support +Inf, -Inf and NaN (documented here: http://docs.oracle.com/cd/E13085_01/doc/timesten.1121/e13070/types.htm), why does this not work correctly through EF?  (NaN does work actually, its just the infinities that have this problem).

      Is there a way to "fix-up" the result before I save it such that it gets stored as +Inf in Oracle?