Posted by safren on 02/15/06 19:15

1. In c++ (or any other language) I'm copying bit by bit to a double
variable dob_var.
2. If I copy dob_var to dob_var2 which is also a double I do not lose
information.
3. This implies that the same would be when copying to a data base, but
when I write dob_var to an sql server float field (same size as a
double variable = 64 bit) it sometimes loses information. e.g. when
value=1E-320.
It is true that SQL float is valid until 1E-308 but so is a C++ double.

Any comments?

[Back to original message]