SMath Function 3

[dik]

Does anyone have a bit of code for a function to confirm that two numbers are approximately equal to each other (using a stipulated difference say 5%?) I can do it if both the numbers are +ve or both are -ve, but it fails at values near 0, where one is +ve and one is -ve. Any help appreciated.

Rather than think climate change and the corona virus as science, think of it as the wrath of God. Feel any better?

-Dik

 

[IDS]

An Excel formula that works is:
=IF(A1-A2=0,1,ABS(A1-A2)/MAX(ABS(A1),ABS(A2))<Tol)
where Tol is a named range with the tolerance.

It should be pretty easy to convert that to SMath.

Doug Jenkins
Interactive Design Services

http://newtonexcelbach.wordpress.com/

 

[dik]

Thanks, Doug... will take a look at it.

Rather than think climate change and the corona virus as science, think of it as the wrath of God. Feel any better?

-Dik

 

[dik]

Rather than think climate change and the corona virus as science, think of it as the wrath of God. Feel any better?

-Dik

 

[MintJulep]

That could be simplified a lot of you embraced Boolean.

 

[dik]

can you give me a hint?

Rather than think climate change and the corona virus as science, think of it as the wrath of God. Feel any better?

-Dik

 

[IDS]

That could be simplified a lot of you embraced Boolean.

I think as written is more readable, but if the function is going to be called from other routines it would certainly make things simpler to return a boolean True/False, or 1/0, rather than a text string.

One general question is whether checking the % difference compared to the larger value is always appropriate. Numbers of different sign will always differ by more than 100%, so for instance 1E-50 and -1E-50 will be found not equal.

Perhaps there should be an absolute value that will be treated as zero, or a value that will be treated as the reference if both values are smaller.

Doug Jenkins
Interactive Design Services

http://newtonexcelbach.wordpress.com/

 

[dik]

...not called from other routines, but as a notification that the two answers are 'correct' or comparable.

Rather than think climate change and the corona virus as science, think of it as the wrath of God. Feel any better?

-Dik

 

[MintJulep]

 

[IDS]

MintJulep - but what if a and b are both zero?

Doug Jenkins
Interactive Design Services

http://newtonexcelbach.wordpress.com/

 

[electricpete]

error... error... does not compute!

=====================================
(2B)+(2B)' ?

 

[IRstuff]

Mathcad version

TTFN (ta ta for now)
I can do absolutely anything. I'm an expert!

https://www.youtube.com/watch?v=BKorP55Aqvg

faq731-376 forum1529 Entire Forum list

http://www.eng-tips.com/forumlist.cfm

 

[MintJulep]

 

[dik]

Thanks, IRS... just what I wanted... I just found out that you can Windows paste directly into SMath...

Rather than think climate change and the corona virus as science, think of it as the wrath of God. Feel any better?

-Dik

 

[MintJulep]

To Doug's point about values straddling zero, this is a much easier question to answer if tolerance is a number rather than a percentage.

 

[IDS]

MintJulep - OK, that's neater

Excel version:
=ABS(A1-A2) <= (Tol * MAX(ABS(A1),ABS(A2)))

I still think we need a check for two numbers very close to zero though.

Doug Jenkins
Interactive Design Services

http://newtonexcelbach.wordpress.com/

 

[dik]

Improved... I had modified the original in case the second arg was a zero... Fixt...

and in the program... the dimensions are all the same because I just 'cut and pasted' the variables and changed the name.

Rather than think climate change and the corona virus as science, think of it as the wrath of God. Feel any better?

-Dik