What does df.dropna() do by default?
df.dropna()
It replaces every NaN with zero.
It returns rows only when none of their values are missing.
It always changes the original df directly.
It removes only the age and city columns.