pandas missing values cheat sheet

Find, drop and fill NaN / None, plus how nulls behave in groupby and comparisons.

Finding

Count nulls per column

df.isna().sum()

Rows with any null

df[df.isna().any(axis=1)]

Drop

df.dropna(subset=["name"])

Fill with a value or statistic

df["salary"] = df["salary"].fillna(df["salary"].median())

Fill within groups

df["salary"] = df.groupby("dept")["salary"].transform(lambda s: s.fillna(s.mean()))

NaN is never equal to NaN

np.nan == np.nan   # False
df["salary"].isna() # use this instead

groupby drops null keys

df.groupby("name", dropna=False).size()

python · editable

loading editor…

Ready for interview-level questions? Pro unlocks the full company problem set.