Sorting simple মনে হলেও time-series problem-এ critical।

df.sort_values("sales")

Descending:

df.sort_values(
    "sales",
    ascending=False
)

Multiple columns:

df.sort_values(
    ["user_id", "date"],
    ascending=[True, True]
)

কেন important?

Operations যেমন:

Lag
Rolling average
Cumulative sum
Previous order
Next order

সবগুলোর আগে correct ordering দরকার।