Adding RoundUp #194
Adding RoundUp #194
Conversation
|
Thanks, @da0x for your contribution. I will review ur PR tomorrow. I have to communicate with Jason (second contributor) as I know he has already started the implementation of RoundUp/Down methods. |
decimal.go
Outdated
| // (1.52).RoundUp(1) => 1.6 | ||
| // (1.1001).RoundUp(2) => 1.11 | ||
| func (d Decimal) RoundUp(places int32) Decimal { | ||
| factor := NewFromFloat(10).Pow(NewFromInt(int64(places))) |
There was a problem hiding this comment.
Decimal Pow is not a high-performance method. I have benchmarked them before but not have results here. You can try to benchmark it.
How about
factor := NewFromFloat(math.Pow(10, float64(-places))There was a problem hiding this comment.
Thank you! Great suggestion, I'll try it out.
There was a problem hiding this comment.
using math.Pow:
BenchmarkDecimal_RoundUp-8 799142 1522 ns/op
using Decimal.Pow
BenchmarkDecimal_RoundUp-8 255872 4411 ns/op
I'll update the PR with your suggestion.
|
I was thinking for a bit about your implementation and I guess there is still a place for improvement. As we decided with Jason to include RoundUp/Down methods in the 1.3 release I would reiterate the implementation to be more performant. Thanks a lot for your contribution and interest in decimal lib! :3 |
|
I agree, N*10^n can be easily optimized by shifting the decimal point instead of multiply/divide. I'm ok with this implementation though since performance isn't an issue for me. You can accept this PR and later replace it with your implementation or just keep it as a reference for others until yours is deployed. Up to you! Thanks for building this library guys. very useful. |
|
Related with this issue #191 |
|
I apologize I still did not create PR with my solution. I had a lot of things to do today. I promise I will do it tomorrow, I will keep you posted :)) |
|
Resolved: #196 (comment) |
Adding round up. Rounds decimal towards positive infinite.