Further improvements to SkillTable and SkillArray

[jsmariegaard]

No description provided.

[ecomodeller]

I am looking at the multivariable comparison, where not all combinations of model, observation and variable exist.

Does this behaviour make sense ,i.e selecting a list of columns or a single column?

In the first case with [["rmse"]] I don't get any missing values, whiile using ["rmse"] I do.

>>> s[["rmse"]].to_dataframe()
                                                rmse
model observation variable                         
SW_1  EPL_Hm0     Significant wave height  0.223597
      F16_wind    Wind speed               2.775330
      HKNA_Hm0    Significant wave height  0.351964
      HKNA_wind   Wind speed               1.276179
      c2_Hm0      Significant wave height  0.378361
      c2_wind     Wind speed               0.698317
SW_2  EPL_Hm0     Significant wave height  0.223597
      F16_wind    Wind speed               2.775330
      HKNA_Hm0    Significant wave height  0.351964
      HKNA_wind   Wind speed               1.276179
      c2_Hm0      Significant wave height  0.378361
      c2_wind     Wind speed               0.698317
>>> s["rmse"].to_dataframe()
                                               rmse
model observation variable                         
SW_1  EPL_Hm0     Significant wave height  0.223597
                  Wind speed                    NaN
      F16_wind    Significant wave height       NaN
                  Wind speed               2.775330
      HKNA_Hm0    Significant wave height  0.351964
                  Wind speed                    NaN
      HKNA_wind   Significant wave height       NaN
                  Wind speed               1.276179
      c2_Hm0      Significant wave height  0.378361
                  Wind speed                    NaN
      c2_wind     Significant wave height       NaN
                  Wind speed               0.698317
SW_2  EPL_Hm0     Significant wave height  0.223597
                  Wind speed                    NaN
      F16_wind    Significant wave height       NaN
                  Wind speed               2.775330
      HKNA_Hm0    Significant wave height  0.351964
                  Wind speed                    NaN
      HKNA_wind   Significant wave height       NaN
                  Wind speed               1.276179
      c2_Hm0      Significant wave height  0.378361
                  Wind speed                    NaN
      c2_wind     Significant wave height       NaN
                  Wind speed               0.698317

[jsmariegaard]

I am looking at the multivariable comparison, where not all combinations of model, observation and variable exist.

Does this behaviour make sense ,i.e selecting a list of columns or a single column?

In the first case with [["rmse"]] I don't get any missing values, whiile using ["rmse"] I do.

>>> s[["rmse"]].to_dataframe()
                                                rmse
model observation variable                         
SW_1  EPL_Hm0     Significant wave height  0.223597
      F16_wind    Wind speed               2.775330
      HKNA_Hm0    Significant wave height  0.351964
      HKNA_wind   Wind speed               1.276179
      c2_Hm0      Significant wave height  0.378361
      c2_wind     Wind speed               0.698317
SW_2  EPL_Hm0     Significant wave height  0.223597
      F16_wind    Wind speed               2.775330
      HKNA_Hm0    Significant wave height  0.351964
      HKNA_wind   Wind speed               1.276179
      c2_Hm0      Significant wave height  0.378361
      c2_wind     Wind speed               0.698317
>>> s["rmse"].to_dataframe()
                                               rmse
model observation variable                         
SW_1  EPL_Hm0     Significant wave height  0.223597
                  Wind speed                    NaN
      F16_wind    Significant wave height       NaN
                  Wind speed               2.775330
      HKNA_Hm0    Significant wave height  0.351964
                  Wind speed                    NaN
      HKNA_wind   Significant wave height       NaN
                  Wind speed               1.276179
      c2_Hm0      Significant wave height  0.378361
                  Wind speed                    NaN
      c2_wind     Significant wave height       NaN
                  Wind speed               0.698317
SW_2  EPL_Hm0     Significant wave height  0.223597
                  Wind speed                    NaN
      F16_wind    Significant wave height       NaN
                  Wind speed               2.775330
      HKNA_Hm0    Significant wave height  0.351964
                  Wind speed                    NaN
      HKNA_wind   Significant wave height       NaN
                  Wind speed               1.276179
      c2_Hm0      Significant wave height  0.378361
                  Wind speed                    NaN
      c2_wind     Significant wave height       NaN
                  Wind speed               0.698317

Nice catch - now fixed