How to represent an Oracle decimal number with precision to a Java type properly?

Hi Objects,
I know what you tried to say by the link...
But that's not quite the problem.........

The real problem is the value itself.
I know you answered a question similar to this one but that wasn't pertinent to this one...

I will use the value in the calculations (mostly multiplications)...
I have a few variables that will hold decimals with differing precision and I will use them to calculate
a value and return the value with differing precison...

Do YOU THINK I SHOULD USE BigDecimal for my varialbes so that the calculating variable can have differing precision and the calculated variable can return an arbitrary precision ????????????

Please let me know...

Thanks and regards.
_Esam

the value is fine, its just how the result is diplayed that is your problem

double gives you ~ 15 significant digits
if u need to be more exact than that then go to BigDecimal

Let me get  clarified then..

If I have
  rate = 12.123456789012 (14,12)
  factor = 1.1234567 (8,7)

Now if I want a result as: result = 12.123456 (8,6)

What type of varialbes should I use for rate, factor, and result  ? (What would be best calculation wise) ? (rate and factor are bean variables)

Yes, I will need to make sure that I return the variable result with six (6) digits after the decimal (two before decimal). --- because I will need to use this result to set the value in another column (of a table)
that expect it thus . . i.e. a number as (8,6)

Now value is more important than display (as thought in previous comment).

Please let me know...

Regards.
_Esam...

if you want the resulting value to be *exactly* that then you have to use BigDecimal

Should I use BigDecimal for all my variables ... rate, factor and result ?
Or just the result (this is returning exactly as (8,6)) ?

If I am using only result as BigDecimal, then can my rate and factor stay as double ?

Thanks.
_Esam./

depends on what value you need to be exact.

>>depends on what value you need to be exact.
Not clear about this?

Because this is what I wanted::
If I have
  rate = 12.123456789012 (14,12)
  factor = 1.1234567 (8,7)

Now if I want a result as: result = 12.123456 (8,6)  which is rate*factor

What type of variables should rate, factor and result be?

Thanks.
_Esam.

any decimal value you want to be exact you need to use BigDecimal, what aren't u clear about exactly?

Sorry to be cryptic...

If I have a method as:

public ? getResult (? rate, ? factor){
 ?  result        
  result = rate*factor;
return result;
}

Here in the method, I am passing the rate and factor from two data bean(s).
And returning a result that I will use to update a table record in the database.

Now, in the tables of the database...
I have rate as: 12.123456789012 (14,12)
 factor as: 1.1234567 (8,7)

and the resulting value as: 12.123456 (8,6)

So, what type of variables should I use for my bean(s) to represent rate, and factor ?
What variable shoud my result be (if I want exacly as (8,6) ?

Hope I am clear ..

Thanks.
_Esam.

I can't answer that question, its your decision :)

> What variable shoud my result be (if I want exacly as (8,6) ?

but whats the point of having the result value exactly 8.6 if the value you use to calculate it are approximations.

Well,
I will use this result to add to a value in a database table that has the format at (8,6) being Oracle Number.

Now say, if I had rate as : 12.123456789012 (14,12) and
 factor as: 1.1234567 (8,7) both being double(s)

And I got the result as: resiult = rate*factor

Can I assume that I can add this result (result) to the database table column having the format as (8,6)

Because I haven't tested how rate*factor would return if all were double(s).

It's not that I want exactly (8,6) .. but I want to make sure that if I had to inert a result that I insert it as
having the value (8,6)

Am I clear ?

Let me know...

Regards,
_Esam.

> 12.123456789012 (14,12)

if its a double then its *not* exactly 14,12
similiar for factor

> but I want to make sure that if I had to inert a result that I insert it as
having the value (8,6)

if thats the db column type then how would it store it as anything other than 8,6

>>>
> 12.123456789012 (14,12)

if its a double then its *not* exactly 14,12
similiar for factor

How's that so?

>>if thats the db column type then how would it store it as anything other than 8,6
That's exactly what I was trying to understand.

Then I must make sure that the db column type matches the value I pass to use and that's why I needed
to make sure that the result I get is as (8,6) in terms of value at the end (extracted from precision).

Because depending on the value of rate*factor , the result could be ?

And I wanted to make sure at the end the value of result is as (8,6)
The result might have some different precision in between...

Please follow this post and let me know the Qs I had...

Thanks.
_Esam

> How's that so?

floating point numbers are an approximation by definition.

>>floating point numbers are an approximation by definition.
To my understanding .. any type is an approximation..

But this is not the isseu...here...
issue here is how best to get the value(s)...

Becasue if I wanted to have a value with precision of at least 12 after the decimal the I should not use
float, I should at leaset use double...

Let me get this clear...

If I have a BigDecimal as 12.123456789012 ....
Then if I need to use a value as 12.123456 from this BigDecimal as float what I do???

That's the q ?

Thanks.
_Esam

float f = bd.floatValue();

but then you're back to an approximation

may help:

http://mindprod.com/jgloss/floatingpoint.html

Just a though:

  If You want to get the greatest possible precission, You should use BigDecimal, the only drawback for doing that, may be the greater resource (memory and time) usage while calculating the result. If the latter is not the problem, why just not use the BigDecimal type?

this link may be interesting for you
http://forum.java.sun.com/thread.jspa?threadID=666149&start=15&tstart=0

All comes to this point again:
If I have double(s)
 double  rate = 12.123456789012 (14,12)
  double factor = 1.1234567 (8,7)

Now if I want a result as: double result = 12.123456 (8,6)

Since double is giving me precision for 15-16 so I am fine with the result.
Now to put the result back to another table column I would like to have it as (8,6) - total 8 digits, 6 after decimal.

Now I really don't need to worry about BigDecimal at this point as I understand or not?

I just want the result with some rounding mechanism to 6 digits after the decimal !

How I do that?

Thanks.
_Esam.

if your database is storing it as 8.6 then thats what its going to store isn't it :)
or you can use what I originally suggested to convert it to a string with two decimal paces.

rounding is another option

http://www.rgagnon.com/javadetails/java-0016.html