[Python] Compute cumulative mean & std on polars dataframe (using over)

I want to compute the cumulative mean & std on a polars dataframe column.

For the mean I tried this:

import polars as pl

df = pl.DataFrame({
'value': [4, 6, 8, 11, 5, 6, 8, 15],
'class': ['A', 'A', 'B', 'A', 'B', 'A', 'B', 'B']
})

df.with_columns(cum_mean=pl.col('value').cum_sum().over('class')
/ pl.int_range(pl.len()).add(1).over('class'))


which correctly gives

shape: (8, 3)
┌───────┬───────┬──────────┐
│ value ┆ class ┆ cum_mean │
│ --- ┆ --- ┆ --- │
│ i64 ┆ str ┆ f64 │
╞═══════╪═══════╪══════════╡
│ 4 ┆ A ┆ 4.0 │
│ 6 ┆ A ┆ 5.0 │
│ 8 ┆ B ┆ 8.0 │
│ 11 ┆ A ┆ 7.0 │
│ 5 ┆ B ┆ 6.5 │
│ 6 ┆ A ┆ 6.75 │
│ 8 ┆ B ┆ 7.0 │
│ 15 ┆ B ┆ 9.0 │
└───────┴───────┴──────────┘


However, this seems very clunky, and becomes a little more complicated (and possibly error-prone) for std.

Is there a nicer (possibly built-in) version for computing the cum mean & cum std?

Continue reading...