polars.Expr.rolling#

Expr.rolling(
index_column: str,
*,
period: str | timedelta,
offset: str | timedelta | None = None,
closed: ClosedInterval = 'right',
check_sorted: bool = True,
) → Self[source]#

Create rolling groups based on a time, Int32, or Int64 column.

If you have a time series <t_0, t_1, ..., t_n>, then by default the windows created will be

  • (t_0 - period, t_0]

  • (t_1 - period, t_1]

  • …

  • (t_n - period, t_n]

whereas if you pass a non-default offset, then the windows will be

  • (t_0 + offset, t_0 + offset + period]

  • (t_1 + offset, t_1 + offset + period]

  • …

  • (t_n + offset, t_n + offset + period]

The period and offset arguments are created either from a timedelta, or by using the following string language:

  • 1ns (1 nanosecond)

  • 1us (1 microsecond)

  • 1ms (1 millisecond)

  • 1s (1 second)

  • 1m (1 minute)

  • 1h (1 hour)

  • 1d (1 calendar day)

  • 1w (1 calendar week)

  • 1mo (1 calendar month)

  • 1q (1 calendar quarter)

  • 1y (1 calendar year)

  • 1i (1 index count)

Or combine them: “3d12h4m25s” # 3 days, 12 hours, 4 minutes, and 25 seconds

By “calendar day”, we mean the corresponding time on the next day (which may not be 24 hours, due to daylight savings). Similarly for “calendar week”, “calendar month”, “calendar quarter”, and “calendar year”.

In case of a rolling operation on an integer column, the windows are defined by:

  • “1i” # length 1

  • “10i” # length 10

Parameters:
index_column

Column used to group based on the time window. Often of type Date/Datetime. This column must be sorted in ascending order. In case of a rolling group by on indices, dtype needs to be one of {Int32, Int64}. Note that Int32 gets temporarily cast to Int64, so if performance matters use an Int64 column.

period

length of the window - must be non-negative

offset

offset of the window. Default is -period

closed{‘right’, ‘left’, ‘both’, ‘none’}

Define which sides of the temporal interval are closed (inclusive).

check_sorted

When the by argument is given, polars can not check sortedness by the metadata and has to do a full scan on the index column to verify data is sorted. This is expensive. If you are sure the data within the by groups is sorted, you can set this to False. Doing so incorrectly will lead to incorrect output

Examples

>>> dates = [
...     "2020-01-01 13:45:48",
...     "2020-01-01 16:42:13",
...     "2020-01-01 16:45:09",
...     "2020-01-02 18:12:48",
...     "2020-01-03 19:45:32",
...     "2020-01-08 23:16:43",
... ]
>>> df = pl.DataFrame({"dt": dates, "a": [3, 7, 5, 9, 2, 1]}).with_columns(
...     pl.col("dt").str.strptime(pl.Datetime).set_sorted()
... )
>>> df.with_columns(
...     sum_a=pl.sum("a").rolling(index_column="dt", period="2d"),
...     min_a=pl.min("a").rolling(index_column="dt", period="2d"),
...     max_a=pl.max("a").rolling(index_column="dt", period="2d"),
... )
shape: (6, 5)
┌─────────────────────┬─────┬───────┬───────┬───────┐
│ dt                  ┆ a   ┆ sum_a ┆ min_a ┆ max_a │
│ ---                 ┆ --- ┆ ---   ┆ ---   ┆ ---   │
│ datetime[μs]        ┆ i64 ┆ i64   ┆ i64   ┆ i64   │
╞═════════════════════╪═════╪═══════╪═══════╪═══════╡
│ 2020-01-01 13:45:48 ┆ 3   ┆ 3     ┆ 3     ┆ 3     │
│ 2020-01-01 16:42:13 ┆ 7   ┆ 10    ┆ 3     ┆ 7     │
│ 2020-01-01 16:45:09 ┆ 5   ┆ 15    ┆ 3     ┆ 7     │
│ 2020-01-02 18:12:48 ┆ 9   ┆ 24    ┆ 3     ┆ 9     │
│ 2020-01-03 19:45:32 ┆ 2   ┆ 11    ┆ 2     ┆ 9     │
│ 2020-01-08 23:16:43 ┆ 1   ┆ 1     ┆ 1     ┆ 1     │
└─────────────────────┴─────┴───────┴───────┴───────┘