Hi, I have bit of a problem with writing unit tests for the new FORECAST.ETS functions (https://bugs.documentfoundation.org/show_bug.cgi?id=94635). As the functions return forecasts and predicted confidence intervals, the returned values are not mathematically exact. Excel2016 returns slightly different values than Calc does, because of differences in the way coefficients are calculated, internal thresholds for error levels and/or accuracy, etc.
So it seems that all I can do is to check handling of (in)correct argument values by checking if the functions return an error or not. I could add something like result should be x plus or minus y%, but that doesn't attract me. I cannot use an xlsx document to test import as well because Excel only supports about half of the new functions (and the results differ anyway). Any suggestions on how to proceed here? Winfried
_______________________________________________ LibreOffice mailing list LibreOffice@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/libreoffice
