对于 Java 中的大数，double 与 long

Question

I'm just starting to learn to code and this might be a very simple question but I have seen double used for numbers that are larger than int can hold. If I understand correctly, double is less precise than using long might be.

So if I have a number larger than int can hold, would it be best to use double or long? In what cases is one preferred over the other? What is best practice for this?

(Say I would like to store a variable with Earth's population. Is double or long preferred?)

Answer 1

tl;dr

For population of humans, use long primitive or Long class.

Details

The floating-point types trade away accuracy in exchange for speed of execution. These types in Java include float / Float and double / Double .

If working with whole numbers (no fractions):

• For smaller numbers ranging from -2^31 to 2^31-1 (roughly plus or minus 2 billion, use int / Integer . Needs 32-bits for content.
• For larger numbers, use long / Long . Needs 64-bits for content.
• For extremely large numbers, use BigInteger .

If working with fractional numbers (not integers):

• For numbers between (2-2^-23) * 2^127 and 2^-149 where you do not care about accuracy, use float / Float . Needs 32-bits for content.
• For larger/smaller numbers where you do not care about accuracy, use double / Double . Needs 64-bits for content.
• For more extreme numbers, use BigDecimal .
• If you care about accuracy (such as money matters), use BigDecimal .

So for the earth's population of humans, use long / Long .