pandas and NumPy 3 cheatsheet

pandas and NumPy 3: Cleaning and transforming

11 common pandas and NumPy 3 tasks.

Count missing values

Count the missing values in each column.

df.isna().sum()

Remove rows that have a missing value in the given columns.

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

Replace missing values in a column with a default.

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

Keep the first row for each email.

df = df.drop_duplicates(subset=["email"])

Rename some columns with a mapping.

df = df.rename(columns={"dob": "date_of_birth"})

Remove columns you don't need.

df = df.drop(columns=["temp", "notes"])

Convert a column to another type.

df["age"] = df["age"].astype("int64")

Create a column from others. Vectorized operations are much faster than apply.

df = df.assign(total=df["price"] * df["qty"])

Choose one of two values per row.

df["group"] = np.where(df["age"] >= 18, "adult", "minor")

Apply string methods to a whole column.

df["email"] = df["email"].str.strip().str.lower()

Convert text to datetimes, then pull out parts.

df["date"] = pd.to_datetime(df["date"])
df["month"] = df["date"].dt.month