Damon, REAL is a very limited data type. It is extremely susceptible to roundoff errors. The value is stored in only 4 bytes, so you have only 6 digits of precision. DOUBLE is far more accurate, because it uses 8 bytes for storage, has 15 digits of precision, and therefore can be precise with numbers billions of times larger than REAL can handle.
Bill On Fri, May 15, 2009 at 4:05 PM, Gray, Damon <[email protected]> wrote: > Let’s suppose we have a table with a REAL and an INT, and let’s lay it > out left to right as below: > > > > REAL1 INT1 > > .0599 NULL > > .0600 NULL > > .0601 NULL > > .0698 NULL > > .0699 NULL > > .0700 NULL > > .0701 NULL > > > > Now, let’s use the update command to set INT1 to the value of REAL1 * > 10000. What would you expect? Well, I expect something other than what I’m > getting. The results are as follows: > > > > REAL1 INT1 > > .0599 599 > > .0600 599 > > .0601 601 > > .0698 697 > > .0699 698 > > .0700 700 > > .0701 701 > > > > What gives?? > > > > > > wwwww > > ( @ @ ) > > -------oOO---(_)---OOo------- > > Damon J. Gray > > Anvil Corporation Business Services > > (360) 937-0770 > > oooo0 0oooo > > ( ) ( ) > > ----------\ (----) /----------- > > \__) (__/ > > >
