Possible issue with calculation

[seekerkoruth]

code:= 1+1/(4+5)*(1*5)+3.3+3.3+3.3+3.3/3.3*3.3

When I run this code in excel, I get 14.7555. When I run it in expr, I get 14.2.

What accounts for the difference?

however this works: 1.0+1/(4.0+5)*(1*5.0)+3.3+3.3+3.3+3.3/3.3*3.3 yields 14.7555

It looks like you have to marshal the integers into floats by adding .0

Thank you.

[antonmedv]

I think this can be an issue: 1/(4+5)*(1*5)

[seekerkoruth]

How do you propose this can be solved? At the code level or at the end consumer level by adding the .0 hint to a literal numerator or denominator when there is division? However, if both the numerator and denominator are both integer variables then we may have less control.

[antonmedv]

What about float as default for every constants in 1/2 expression?

[seekerkoruth]

What about float as default for every constants in 1/2 expression?

I think this will work. Float as a default for all division operations is a excellent rational solution to this issue as this will cover integer variables as numerators or denominators also and not just constants/literals.

[antonmedv]

@davidklassen what do you think?

[mrxrsd]

There is another issue related to precedence, I don't know if it is the same root cause. The expression "-(1*2)^2" it is expected "-4" but it was returned "4".

[antonmedv]

The expression "-(1*2)^2" it is expected "-4" but it was returned "4".

The correct answer is 4.

[mrxrsd]

Not according to google and wolfram alpha =p

https://www.wolframalpha.com/input/?i=-%2812%29%5E2 https://letmegooglethat.com/?q=-(12)%5E2

[antonmedv]

I guess unary op is higher priority than exponential in expr.

[antonmedv]

Fixed in https://github.com/antonmedv/expr/commit/2aec311b2bcdb9ab5a8efee88c08138dde84a65c