We are celebrating the 10-year anniversary of the Alteryx Community! Learn more and join in on the fun here.
Start Free Trial

Alteryx Designer Desktop Discussions

Find answers, ask questions, and share expertise about Alteryx Designer Desktop and Intelligence Suite.
SOLVED

RawToChar error message when running Logistic Regression Tool.

JohnCavil
5 - Atom

Hey,

 

I'm having trouble running a logistic regression. I keep getting a - rawToChar(serialize(the.object, connection = NULL, ascii = TRUE) - error. I only seem to get this error if i run my full dataset however, which is 6 million rows. If i run a million rows it runs just fine.

 

Below that error i also get a - Info: Linear Regression (122): Tool #4:   long vectors not supported yet: raw.c:68 - error.

 

I've looked into other people having this problem, and the consensus seems to either be a character error of some sort (meaning i have a strange character somewhere in my data) or that it's some sort of size limitation of the data set. I've ruled out the data being wrong. It's all cleaned up and there is no error in the data.

 

According to this post: https://community.alteryx.com/t5/Alteryx-Designer/Linear-Regression-error-only-when-feeding-large-da... it seems to be a problem with the size of the dataset, some sort of limitation within R or something. However a real solution was never really given.

 

Seeing as how i can run this with 1 million rows but not with 6 million it seems likely that this is the error.

 

Can anyone confirm that this is indeed the problem, and if so, how to get around it?

 

Any help would greatly be appreciated 🙂

1 REPLY 1
TrevorS
Alteryx Alumni (Retired)

Hello @JohnCavil 

The message following the Error in rawToChar(serialize(the.object, connection = NULL, ascii = TRUE)) indicates the root cause of the error, and why the Logistic Regression Tool is failing to build a model. The message long vectors not supported yet: raw.c:68 is indicating that in the version of R running the code for the Logistic Regression Model, the vector created containing the model object is too long to be serialized into an object that can be passed back to the Alteryx Engine. Essentially, you are passing too many records for the Logistic Regression Tool to train and return a model in Alteryx. You can read more about Long Vectors in R here: https://stat.ethz.ch/R-manual/R-devel/library/base/html/LongVectors.html as well as some discussion of the error in different contexts here:

https://github.com/ropensci/bold/issues/29
https://stackoverflow.com/questions/24335692/large-matrices-in-r-long-vectors-not-supported-yet


Shoutout to @KyleB  for the above solution!

Community Moderator
Labels
Top Solution Authors