Getting the NaN rows from pandas.dropna-CodePudding

I am using dropna to get rid of the NaN values, but instead of just dropping them i want to get a new table where those rows are saved. That's to say from the current code:

df_weight.dropna(subset = ["age"], inplace=True)
df_weight.dropna(subset = ["height"], inplace=True)
df_weight.dropna(subset = ["weight"], inplace=True)
df_weight

i want to save the rows that are droppen in the line df_weight.dropna(subset = ["weight"], inplace=True). I think that dropna does not have a return value, so there is any work around to archive this?

EDIT: my db comes from

CodePudding user response：

you can try as follows. if you share the DF, it will be easier to reproduce and provide the working solution

its an idea or direction

df_weight.isna()['age']
df_weight.isna()['height']
df_weight.isna()['weight']

CodePudding user response：

You could use

dropped_rows_df =  df_weight[df_weight.isna(axis=0, how='any', subset=['age','height','weight'])]