Why are some of my numeric features not being recognized as 'numeric' types AND why can't I reclassify them?

I can't share my CSV here but I can assure you those features are indeed numeric (I use the same dataset in R studio without any problem).

I have found that the presence of "NULL" values was one trigger (would there be a way to tell the importer to treat "NULL" as NA? I couldn't find anything related to this in the documentation...) but once I filtered them out of my CSV, some features still get the wrong type allocated to them.

The screen grab below presents the only two choice the software gives me...

No choice to change the feature type to numeric

How can I properly set the type of an attribute in Orange? And how should I handle the missing values represented as "NULL"?

  • $\begingroup$ Can you just post a fake 1 line CSV file? Or you could try to split the file in half - try to load it, if it doesn't - try again, and so on, until you find a subset that works and one that doesn't. Then you can figure out why exactly. $\endgroup$ – Silent M Aug 9 '18 at 21:12
  • $\begingroup$ Thanks for the comment/edit : I sadly have no explanation on how I fixed this issue other than that I re-installed Orange... $\endgroup$ – Sylvansky Aug 16 '18 at 7:51

I restarted the computer, re-installed Orange and I haven't had any issues since... A bug I guess.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.