Timeseries - Strange behaviour of retime() function

Dear Scilab user,

I’m working on big 10-minutes meteorological data and I’m trying to use the retime function to get “daily means”.

The function has a strange behaviour that I cannot understand.

Here’s an example :

---> r = retime(ts_meteo(1:52600,:), 'daily', nanmean)
r  = 

366x2 timeseries
          Time           Temperature   Humidite 
   ___________________   ___________   _________
                                                
   2023-03-02 00:00:00   3.7777778     75.428889
   2023-03-03 00:00:00   4.8194444     72.328819
   2023-03-04 00:00:00   1.7430556     79.397014
           ...               ...          ...   
   2024-02-28 00:00:00   3.3333333     98.117083
   2024-02-29 00:00:00   6.8194444     76.605556
   2024-03-01 00:00:00   6.7410072     75.556115

retime() function is working well for the first 52600 lines of the matrix but :

---> r = retime(ts_meteo(1:52610,:), 'daily', nanmean)
 r  = 

367x2 timeseries
          Time           Temperature   Humidite 
   ___________________   ___________   _________
                                                
   2023-03-02 00:00:00   3.7777778     75.428889
   2023-03-02 00:10:00   4.8194444     72.328819
   2023-03-02 00:20:00   1.7430556     79.397014
           ...               ...          ...   
   2023-03-04 12:40:00   6.8194444     76.605556
   2023-03-04 12:50:00   6.8724832     74.209128
   2023-03-04 13:00:00   Nan           Nan      

retime() fails with 10 more lines. The time vector is wrong. But the means seems to be the same at least at the beginning.

Below, the detail of the 10 additional lines is shown. However everything looks normal…

--> ts_meteo2 = ts_meteo(52600:52610,:)
 ts_meteo2  = 

11x2 timeseries
          Time           Temperature   Humidite
   ___________________   ___________   ________
                                               
   2024-03-01 23:00:00   8             54.91   
   2024-03-01 23:10:00   9             55.2    
   2024-03-01 23:20:00   8             55.48   
   2024-03-01 23:30:00   9             55.7    
   2024-03-01 23:40:00   8             55.76   
   2024-03-01 23:50:00   9             55.46   
   2024-03-02 00:00:00   8             55.69   
   2024-03-02 00:10:00   9             55.06   
   2024-03-02 00:20:00   9             55.26   
   2024-03-02 00:30:00   9             55.52   
   2024-03-02 00:40:00   9             55.73   

The problem occurs just after end of february. Is this due to the bisextile year ?
Or, is there a limitation of the number of days (366 ?) in the function retime() ?

Thanks a lot for your help.

Regards,

Laurent

Hello Laurent,

Can you share your Time column data with me so that I can test it please?

Hello Adeline,

Please find attached, timeseries : Time / Temperature / Humidite that I work on.

ts_meteo.txt (2.4 MB)

Thank you.

Best regards,
Laurent

Hello Laurent,

I can’t reproduce your issue with Scilab 2025.0.0. Could you test with this version, please ?

Adeline

Hello Adeline,

You are right, the new version of Scilab 2025.0.0 corrected the issue.
Thank you.

Laurent

1 Like