search

rethomics / zeitgebr

Circadian rhythm analysis and visualisation in R
11 stars 7 forks source link

Getting "Error in t/bin_length : non-numeric argument to binary operator" for several functions #12

Closed pubhatta closed 4 years ago

pubhatta commented 4 years ago

Hi,

I've created my behavr table from my metadata and data files, which are structured very similarly to the examples in the behavr tables tutorial. However, I've tried visualizing the data using ggetho - stat_tile_etho() and stat_bar_tile_etho(), and I've tried computing a chi square periodogram and each time I get this error:

"Error in t/bin_length : non-numeric argument to binary operator"

I made sure that my 't' column is numeric and the values are in seconds (as shown in the tutorial examples), and I assume 'bin_length' is a variable somewhere within the function's code, so I'm not sure where the 'non-numeric argument' is coming from. I even specified the sampling rate in the periodogram method, but it didn't make a difference. The only code I was able to run was when I specifically ran the "chi_sq_periodogram" function on just the activity column, but this only gives me one observation for each period, rather than observations for each period in each subject's data.

My data, metadata and code files are in the link below: https://github.com/pubhatta/zeitgebr_issue

Any help fixing this issue is appreciated!

qgeissmann commented 4 years ago

Hey, thanks for reporting. just sent you a pull request (https://github.com/pubhatta/zeitgebr_issue/pull/1). Main issue is time should be named t, not time. Also other issues with some variable names such as uid that are not there (that you need to make yourself to describe individuals in a more informative way than a number). I hope it helps. If all done, feel free to close

pubhatta commented 4 years ago

Thanks for the response. Regarding uid , does this make a difference in terms of the code working? I can see how a more descriptive id is beneficial for being more informative, but since my data came formatted with numbers for id's and the tutorial recommends assigning numeric uid's to help with plotting, I figured I could skip that step because my id's are already numeric. Why would this affect the functioning of the code?

pubhatta commented 4 years ago

Never mind, I just realized that uid is used in the facet_wrap() argument! Must have missed that earlier.