addition accuracy

[Jason Hocker]

Why does this return the value 55296.19000000006? How can I get 55296.19?

Double d =  new Double(121145.85);
Double dd = new Double(118338.67);
Double e = d.doubleValue() + dd.doubleValue();

[Keith Lynn]

Floating point numbers usually don't have exact representations. When you print out a double, you can use the DecimalFormat class.

[Joe Vahabzadeh]

I think the only case where you're guaranteed 100% accuracy is when the fractional parts of the values involved can be represented by sums of negative powers of 2.

[Ulf Dittmer]

Point #20 in the JavaBeginnersFaq covers this.