Grouping and aggregation Updated November 2026

cumsum

df["col"].cumsum() returns the running total: each row holds the sum of all values up to and including that row. df.groupby("key")["col"].cumsum() restarts the total for each group while keeping one value per original row.

Syntax

Series.cumsum(skipna=True)   |   Series.cumprod()   |   Series.cummax()   |   Series.cummin()

Examples

import pandas as pd

df = pd.DataFrame({"month": ["Jan", "Feb", "Mar"], "sales": [100, 120, 90]})
df["ytd"] = df["sales"].cumsum()
df

Result: month sales ytd 0 Jan 100 100 1 Feb 120 220 2 Mar 90 310

g = pd.DataFrame({"id": ["a", "a", "b", "b"], "v": [1, 2, 10, 20]})
g["run"] = g.groupby("id")["v"].cumsum()
g

Result: id v run 0 a 1 1 1 a 2 3 2 b 10 10 3 b 20 30

(df["sales"].cumsum() / df["sales"].sum() * 100).round(1).tolist()

Result: [32.3, 71.0, 100.0]

Cumulative percentage of the total.

Common mistakes

The running total is in the wrong order: cumsum follows row order, so sort by date first.
NaN breaks the sequence: with skipna=True NaN rows stay NaN and the total continues after; fill them with 0 if they mean zero.
Using sum when you want cumsum: sum returns a single number, cumsum one number per row.

Practice cumsum

These drills check your answer on live data, free, no account needed.

Running totalReturn the running total of `signups`, day by day.

Questions

How do I reset the cumulative sum when a condition is met?

Create a group id that increments at each reset, for example grp = cond.cumsum(), then df.groupby(grp)["v"].cumsum().

How do I count rows cumulatively per group?

df.groupby("id").cumcount() + 1 numbers the rows 1, 2, 3 within each group.

rollingHow to compute a moving average in pandas with rolling: window size, min_periods to avoid leading NaN, centered windows, rolling sum and std, time windows.
shift and diffHow to use shift and diff in pandas to get the previous or next row, compute the change and percent change between periods, and lag within groups.
groupbyHow groupby works in pandas: group by one or more columns, apply sum, mean, count or size, get a flat table with as_index=False, and iterate over groups.
transformHow groupby transform works in pandas: broadcast a group sum or mean to each row, share of group total, group-wise normalization, and transform versus agg.