The description of the pull request/patch is wrong then, because DB1_DOUBLE is a floating point, not an integer.
Eventually you can add support for both DB1_BIGINT and DB1_DOUBLE. However, few remarks:
* use if-else, to be more optimal and avoid unnecessary evaluations of conditions, with check for DB1_INT being the first * for DB1_BIGINT and DB1_DOUBLE do cast to (int), otherwise there will be warnings on strict compilers