Is there an ability to query the cumulative normal distribution from the Z score - much like =NORM.S.DIST(2,TRUE) does in Excel?
Either an In-DB and/or function would be perfect!
If you are happy using a third party library for Alteryx, I have a NormDist function in my formula addins (https://github.com/jdunkerley/AlteryxFormulaAddOns/releases/tag/v1.2)
YnEoS.
Out of the box, the function is not explicitly present in Alteryx. One can however use the R tool and create a macro that delivers the function. Alternatively, the function can be configured using standard tools.
@DylanB created a lnorm function as an answer to a previous post (search for "distribution").
https://community.alteryx.com/t5/Advanced-Analytics/Lognormal-Distribution-Cumulative-Function/m-p/23005
Creation or modification of this macro could be all that you need.
Cheers,
Mark
Hi,
Thanks for sharing!
I tried installing it and can get the string/misc/date utils to work but not the math utils.
Anything special to be done regarding the math utils?
Thanks,
Hadrien
I've installed the math functions, but i doesn't work. The files are correctly copied, but the formula tool cant find the normdist function. I've tryed to follow the steps in your pdf specification, but it still doesnt work.
Do you have any idea if theres something different in the new 2018.3?
Hello, I was able to successfully install the files, and the NormDist function works when I specify the X, mean, and StdDev within the formula node. However, when I try to link up to a text file with the same inputs, it gives me an error. I think the error is specifically for the standard deviation value, as it runs successfully up until I use the text file as an input for the StdDev field.
This is working:
This is not working:
Error:
"Formula (16) Parse Error at char(0): Error in parsing a formula: Error in function boost::math::cdf(const normal_distribution<double>&, double): Random variate x is -nan(ind), but must be finite! (Expression #2)"
@lross492 Thanks, I'll take a look and get back to you!
@lross492 I looked at our source files and the latest version shouldn't have the issue you're seeing. We also ran into the same issue as you and applied a fix. You'll just need to dowload the latest dll and xml files from github: https://github.com/jdunkerley/AlteryxFormulaAddOns. Then you'll need to replace your current abacus dll and xmls with the new ones. The files are in the %AlteryxBinDirectory%\RuntimeData\FormulaAddIn folder, so something like C:\Program Files\Alteryx\bin\RuntimeData\FormulaAddIn for the standard install.
Thanks Patrick.
When I click on the AlteryxAbacus.dll file, there is an option to download which I have done. I am not seeing a similar download options for the 4 .xml files.
How would you suggest replacing the .xmls?
Update: I just manually created the .xmls using the raw data from the link. Looks to be working!
@lross492 Awesome!
Hi Patrick,
I tried to installed by copying all the dll and xml files into my Formula Add In as the install.bat and Analytics App didnt work for me. However, it still gave me the error" Building custom formula functions using a DLL is not available on AMP yet (Expression #2). Can you advise my case ?
Hi @linhtran710! The NormDist (and all other C++ functions) are not yet compatible with the AMP engine. For any workflow that you want to use NormDist, you'll need to ensure that AMP is off (uncheck the box). AMP may support C++ functions in 23.2 according to the latest update back in March.
I unchecked the AMP, but the formula only works with syntax NORMDIST(X). When I use the full formula NORMDIST(X, 0, 1, 0), it had unknown var type error at NORMDIST. Is there anything I could do ?
@linhtran710 Sorry for the delay, what version of alteryx are you using? 23.1.1.123 had a bug at launch (TDSL-4785), but they fixed it via Patch1 (23.1.1.200). If you're on 23.1.1.123 (and can't/don't want to patch), you can multiply arguments by 1.0 like normdist(x,0*1.0,1*1.0,0*1.0). I'm not sure if that last one needs multiplied, but definitely the 1st and 2nd numbers.
