Eclipse Community Forums: Epsilon » EOL Real number addition overflow

Help

Home

Home » Modeling » Epsilon » EOL Real number addition overflow

Show: Today's Messages :: Show Polls :: Message Navigator

EOL Real number addition overflow [message #1803678]

Wed, 06 March 2019 16:12

Bas Har

Messages: 18
Registered: May 2017

Junior Member

I am trying to perform an addition on real number in an EOL script. Some of the calculations keep overflowing.
Does there exist an approach to handle this overflow?

Example:
3.0E-8 + 4.0E-7
The result should be 4.3e-7
But EOL returns 4.2999999999999996E-7

Appreciate any assistance.
Thank you.

Report message to a moderator

Re: EOL Real number addition overflow [message #1803688 is a reply to message #1803678]

Wed, 06 March 2019 19:03

Antonio Garcia-Dominguez

Messages: 594
Registered: January 2010
Location: Birmingham, UK

Senior Member

We are using standard Java (IEEE 754) floats / doubles, no special magic :-). Remember that IEEE 754 floating point numbers are always approximate: some numbers are not directly representable. If you need something with arbitrary precision, BigDecimal may be a better choice.

Here is a quick StackOverflow article which discusses the issue:

https://stackoverflow.com/questions/3413448/double-vs-bigdecimal

Report message to a moderator

Previous Topic:	Problem compiling RCP using tycho and epsilon
Next Topic:	Need help getting started with EUnit

Goto Forum:

-=] Back to Top [=-

[ Syndicate this forum (XML) ] [

]

Current Time: Thu Apr 25 17:14:15 GMT 2024

.:: Contact :: Home ::.

Breadcrumbs

Sign up to our Newsletter