[Python] How to take head(n) and tail(n) in one group_by with polars

Here is a sample dataframe, I want to get head(n) rows and tail(n) rows in everyday which is a group_by("date").agg() with polars. I know it could use two group_by to get head and tail part and then concat them.

df = pl.from_repr("""
┌────────────┬────────┬──────────────────┐
│ date ┆ symbol ┆ ts_dom2secdom_op │
│ --- ┆ --- ┆ --- │
│ date ┆ str ┆ f64 │
╞════════════╪════════╪══════════════════╡
│ 2000-01-04 ┆ AL ┆ -0.119165 │
│ 2000-01-04 ┆ RU ┆ 0.256691 │
│ 2000-01-05 ┆ AL ┆ -0.126549 │
│ 2000-01-05 ┆ RU ┆ 0.1851 │
│ 2000-01-06 ┆ CU ┆ -0.121354 │
│ 2000-01-06 ┆ RU ┆ 0.228452 │
│ 2000-01-07 ┆ AL ┆ -0.126013 │
│ 2000-01-07 ┆ RU ┆ 0.348729 │
│ 2000-01-10 ┆ AL ┆ -0.139447 │
│ 2000-01-10 ┆ RU ┆ 0.263048 │
└────────────┴────────┴──────────────────┘
""")


Is there some trick to do this in one group_by or with_columns or filter?

Continue reading...