[Python] Cannot get dataframe.['col1].median() to ignore nulls

I have pandas dataframe df1 with a 'rating' column. I Need the median of that column. The column contains the string 'Not Given' for missing values. I have replaced the 'Not given' values Numpy's np.nan and None using the Line:

df1.loc[df1['rating'] == 'Not given', 'rating'] = np.nan`
-OR-
df1.loc[df1['rating'] == 'Not given', 'rating'] = None`
and verified the results. Then tried to get the median of df1['rating'] using this line:
mrat_y =df1['rating'].median( skipna = True)



The error message I get using either null substitute is:

{File "C:\Users\fnaso\PycharmProjects\pythonProject\.venv\Lib\site-packages\pandas\core\nanops.py", line 787, in nanmedian
raise TypeError(f"Cannot convert {values} to numeric")
TypeError: Cannot convert [nan nan '5' ... nan '5' nan] to numeric
}



I am using Pycharm community edition with the latest Python version.

So either the replacements are not valid Nulls or median( skipna = True) is working as I expect or (most likely) I am missing something with my newbie head up my ass.

This works with theW3Schools_Tryit_Editor.

I was expecting to find the median of a panda dataframe column that has float8 and Null values.

Thank you all in advance. I hope some day that can contribute to this fine forum.

Continue reading...