float vs double (in Java)

Question

Is there ever a case where these two methods would return different values given the same inputs?

int compare1(float a, float b)
{
    return Double.compare(a, b);
}

int compare2(float a, float b)
{
    return Float.compare(a, b);
}

By the same token, is it true (or false) that any number storable in a java's Float can be stored in a java's Double without losing any precision?

Thanks

SLaks · Accepted Answer · 2013-01-21T18:25:37.973

8

~~Yes; casting doubles to floats can yield different results.~~

If the difference between a and b is too small to show up in a float, compare2() will return 0 whereas compare1() would not.

You just edited the question to reverse what you were asking. The new answer is:

I'm almost certain that they will always be the same.

edited Jan 21 '13 at 18:25

answered Jan 21 '13 at 18:01

SLaks

868,454
176
1,908
1,964

What about my question about Float being a subset of Double? – user1508893 Jan 21 '13 at 18:03
@user1508893: I'm reasonably certain that it is, but I don't know for sure. – SLaks Jan 21 '13 at 18:03
Actually, I made some mistake in the question. Just updated it. Sorry! – user1508893 Jan 21 '13 at 18:19

score 4 · Answer 2 · edited May 23 '17 at 10:24

4

Every float can be represented exactly as a double. From this it follows that the two functions always return the same result.

edited May 23 '17 at 10:24

Community

1
1

answered Jan 21 '13 at 18:27

NPE

486,780
108
951
1,012

score 1 · Answer 3 · answered Jan 21 '13 at 18:09

1

A double just gives you more bits of precision beyond the decimal place than the float. If all those those extra bits are zero then you have the same value as the float. So yes, all floats can be converted to doubles without losing precision.

answered Jan 21 '13 at 18:09

Eduardo

8,362
6
38
72

1

It also gives you a maximum exponent of 1023 as against 127. – user207421 Jan 21 '13 at 23:07

float vs double (in Java)

3 Answers3