One way is to put the float value into an integer value, which removes the fractional part, and then subtract the integer from the float to get the fractional part.

And then just multiply by 10^(number of digits) and round. So to get 3 digits:
Another way that will always return you all decimals is using Strings:
This will not work, however, if scientific notation (with E, e.g. 4.6E-1) is used. BigDecimal can then help you out, with its toPlainString method.