search

business-science / anomalize

Tidy anomaly detection
https://business-science.github.io/anomalize/
339 stars 61 forks source link

Warning message: `cols` is now required. #43

Open Fredo-XVII opened 4 years ago

Fredo-XVII commented 4 years ago

Hello,

I received the warning below, but search your repo and didn't find any reference to it. Can you add documentation or clarification on this warning? Does it mattter?

Warning message:
`cols` is now required.
Please use `cols = c(anomalies)`

Thanks,

Alfredo

WajdiBenSaad commented 4 years ago

Hello @Fredo-XVII ,

Could you please share a reproducible example (with code & data) so we can test why you are getting this error message.

Thanks

Fredo-XVII commented 4 years ago

Unfortunately, I am using this in a function at work, so I can't share that code or data. I was hoping you would know why I would need to specify "cols = c(anomalies)", or where that would be specified.

mdancho84 commented 4 years ago

It's most likely related to the tidyr v1.0.0 change which requires unnest() to specify cols argument. I thought I fixed this in the last release of anomalize. Try upgrading to the latest CRAN version using install.packages("anomalize"), and see if it goes away. If it persists, I will need a reproducible example to diagnose and resolve.

Fredo-XVII commented 4 years ago

@mdancho84 I see. If that is the case, then I will include that in my unnest() code as well, as I am applying the anomalize functions across hundreds of time series in a nest. So it may not be your package.

Thanks for the reply, I'll close if that resolves my issue.