I need to do calculations on units of time represented in whole and/or tenths of an hour and wonder if BigDecimal is the appropriate datatype for it. From what I gather BigDecimal would be appropriate for currency calculations so am I thinking the same is true in cases where I have to deal with 1.5 + 0.6 type of simple math. Is my thinking correct here?
2
-
A tenth of an hour is 6 minutes, why would you need arbitrary precision for that?– Jon LinCommented Jul 21, 2012 at 4:00
-
Hobbs meter: en.wikipedia.org/wiki/Hobbs_meter– vectorCommented Jul 21, 2012 at 4:04
Add a comment
|
2 Answers
The short answer is yes. If you are performing decimal calculations then BigDecimal is a good class to use to avoid any kind of rounding errors. However, it's probably not the most efficient way. If performance is not an issue then I would say that BigDecimal will serve your purpose.
An alternative would be to perform your calculations in minutes and then divide your answer by 6 to provide the result. In that case you could use an Integer and hold off on the expensive calculation until the end. When you divide by 6 you could cast your answer into a double.
Double answer = 154 / 6.0 //154 represents calculated minutes
answered Jul 25, 2012 at 14:55
-
Performance would not be an issue indeed. All data entry from user end would be in the form of 1.1 or 1.6 etc. I'm afraid anything else would be too much of a burden.– vectorCommented Jul 25, 2012 at 15:22
Because you question mentions that you only need to work with tenths of an hour, it is base you representation on multiples of this smallest time unit, i.e. 6 minutes.
This way it is possible to represent time with an integer data type like, Integer or BigInteger, depending on the maximum time to represent.
You could even write a custom class which wraps the (Big)Integer and provides methods for formatting to string, conversion between other types, etcetera.
-
... possibly, but I'm also trying to keep things as simple as possible.– vectorCommented Jul 21, 2012 at 9:28