> Microsoft's SQL Server doesn't allow any of the special values in > float/double fields. (i.e. positive/negative infinity or NaN). It is > very annoying.
If +INF/-INF are to be stored along with NaNs, and the database server (Derby in our case) doesn't allow it directly, I don't see how one can accomplish this w/o introducing an extra column (of type TINYINT, for instance) which would store a boolean flag to tell you whether your double value is a regular or a special one. -- Andrew ``Bass'' Shcheglov http://www.2ka.mipt.ru/~bass/ http://www.softlogic.ru/