rakesh verma

Greenhorn

Posts: 22

posted 11 years ago

In our project, we are taking care that all the float numbers shuld be of two decimal places only and we r not using something like BigDecimal for this to keep it simple.

So, at all the divisions I am taking care of rounding it to two decimals. Now I found a strange test case, where in adding two numbers with two decimals each resulting in more decimals as below...

S.O.P(56.56+0.56) --- gives 57.120000000000005

S.O.P(56.56f+0.56f) --- gives 57.120003

Ideally it shud just result 57.12. Can I know the reason why it is happening like above.

So, at all the divisions I am taking care of rounding it to two decimals. Now I found a strange test case, where in adding two numbers with two decimals each resulting in more decimals as below...

S.O.P(56.56+0.56) --- gives 57.120000000000005

S.O.P(56.56f+0.56f) --- gives 57.120003

Ideally it shud just result 57.12. Can I know the reason why it is happening like above.

posted 11 years ago

See Some things you should know about floating-point arithmetic.

"We're kind of on the level of crossword puzzle writers... And no one ever goes to them and gives them an award." *~Joe Strummer*

sscce.org

Darryl Failla

Ranch Hand

Posts: 129

posted 11 years ago

Since you're only dealing with two digits all the time, you might want to consider keeping all of your numbers as ints. If it's possible, multiply by 100 and store as an int. Add your ints into a double and display your double divided by 100.

int i1 = 5656;

int i2 = 0056;

double d1 = i1 + i2;

System.out.println(d1/100);

int i1 = 5656;

int i2 = 0056;

double d1 = i1 + i2;

System.out.println(d1/100);

Darryl Failla

Sun Certified Java 2 Programmer

Peter Chase

Ranch Hand

Posts: 1970

posted 11 years ago

This is good advice. And if, as I suspect from the fact you're using 2 decimal places, you are doing financial calculations, it is essential that you use integer arithmetic. Floating point is unacceptably inaccurate.

Not sure about the bit suggesting displaying using a double. This could still go wrong.

[ July 19, 2006: Message edited by: Peter Chase ]

Originally posted by Darryl Failla:

Since you're only dealing with two digits all the time, you might want to consider keeping all of your numbers as ints. If it's possible, multiply by 100 and store as an int.

This is good advice. And if, as I suspect from the fact you're using 2 decimal places, you are doing financial calculations, it is essential that you use integer arithmetic. Floating point is unacceptably inaccurate.

Not sure about the bit suggesting displaying using a double. This could still go wrong.

[ July 19, 2006: Message edited by: Peter Chase ]

Betty Rubble? Well, I would go with Betty... but I'd be thinking of Wilma.