Convert float to rational

[Mellobuck]

Hello,

The TIFF tag header protocol requires a data type I am not really conversant with, the Rational Data type: http://en.wikipedia.org/wiki/Rational_data_type

I need to convert double precision numbers to this data type, e.g. 24.5 needs to be a ratio of two integers: {49,2}. Is there an existing LabVIEW algorithm to perform this operation? All my NI searches keep resulting in Rational ClearCase or Rational Array interpolation. I seem to have found a lot of the theory behind this: http://stackoverflow.com/questions/95727/how-to-convert-floats-to-human-readable-fractions

But, is this implemented in labVIEW already, or should I implement the above-referenced c-code?

Thanks,

Mello

[smercurio_fc]

That implementation is not in LabVIEW. Perhaps you may wish to post a suggestion to add it to a future release of LabVIEW.

If you have C code that will do it then you can call it using Call Library Function Node . You'll either need a DLL or you'll need to create one. Or, you can just convert the code to LabVIEW. 

Message Edited by smercurio_fc on 02-24-2010 11:01 AM

[Mellobuck]

Thank you smercurio, I think I am going to go with the workaround/kluge method of multiplying by a power of ten and truncating:

23.5  => {23500, 1000}

I will add a suggestion that a more elegant solution be made part of the NI math or numeric palettes.

-Mello

[DavidJCrawford]

Hi,

Maybe you can make use of Altenbach's fraction code in the last post in this  link can help you.

cheers

David

Message Edited by David Crawford on 02-24-2010 06:09 PM

[PhillipBrooks · Accepted solution]

This should be possible if you have MathScript available.

 see attached ( LV 8.6.1)

Message Edited by Phillip Brooks on 02-24-2010 02:33 PM

[Mellobuck]

Thanks Phillip!

A Matlab user such as myself should have thought of that method 🙂

[PhillipBrooks]

Glad to be of help. KAPOW!

[smercurio_fc]

---

Phillip Brooks wrote:

This should be possible if you have MathScript available.

---

Neat. Didn't know that function existed in MathScript.

Note that for LV 2009 one requires a separate license for MathScript. There was a big long debate about this when 2009 was released. Every bitched and moaned. It was fun reading. Not so much fun for those who had to shell out more cash for it though.

[PhillipBrooks]

I was curious if there were other ways to do this and found that Scilab combined with the LabVIEW Gateway works quite well. The 'rat' function is available in Scilab; Scilab and the LabVIEW gateway were both easy to install...

http://www.scilab.org/products/other/labview_gateway

Message Edited by Phillip Brooks on 03-04-2010 08:29 AM

[Gajic4x4]

Since LabVIEW has a vi to return the greatest common multiple, it's not hard to make a vi to get to fractions. This snippet solves a similiar problem of converting the floating point number to string whole number plus a fraction.