Dave Cramer wrote:
> I was thinking that parsing the double would be less efficient than
> Byte.parseByte().
>
> The spec suggests that the user would know what they are doing calling
> getByte on a double column.
Yeah, specifically table B-6 says you can use getByte() on pretty much
any numeric SQL type, and (bizarrely) on varchar, char, longvarchar, and
boolean types too.
The spec & javadoc for java.sql.DataTruncation implies we should
generate DataTruncation as a warning if we truncate data on read. I
suppose that loss of precision counts as truncation?
-O