This week's book giveaway is in the Security forum.We're giving away four copies of Penetration Testing Basics and have Ric Messier on-line!See this thread for details.
Win a copy of Penetration Testing Basics this week in the Security forum!

# problem/consult with precision of division with BigDecimals

Manuel Jordan
Ranch Hand
Posts: 131
Hello guys

i have this situation

a and b 2 BigDecimals

a=14.00
b=6.00

for a simple calculator
a/b = 2.3333......
so for humans 14/6 = 2.33 good by infinite

for java 14/6 = 2.34 (a.divide(b,2))
but with (a.divide(b,3)) 2.33
why?

it should be the same?

regards

Jim Yingst
Wanderer
Sheriff
Posts: 18671
The reason for the difference between 2.34 and 2.33 is that when you say a.divide(b,2)), that "2" is the rounding mode - which is BigDecimal.ROUND_CEILING. While "3" is BigDecimal.ROUND_FLOOR. In one case you're telling the computer to round up, and in the other you're telling it to round down. No surprise there. If you want it to round to whichever is closest, use BigDecimal.ROUND_HALF_DOWN or BigDecimal.ROUND_HALF_UP. The difference is very minor. You can examine the JavaDoc to learn the exact differences between rounding modes.

If you're asking why the result is 2.33 rather than 2.33333 or 2.3333333333333 or 2.333333333333333333333333333333333 or whatever, then the answer is that you need to tell it how many digits you want. If you want 10 digits after the decimal point, use a scale of 10:

a.divide(b,10, BigDecimal.ROUND_CEIL) = 2.3333333334
a.divide(b,10, BigDecimal.ROUND_FLOOR) = 2.3333333333

If you want more digits or less, use another number besides 10.

Since you originally did not specify the number of digits you wanted, BigDecimal assumed you wanted the same number as were used for the first BigDecimal in the calculation, a. Since you created this as 14.00, it had two digits after the decimal. That's what was used in your original answers, 2.34 and 2.33.