Create rolling groups based on a date/time or integer column

Description

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\]

Usage

<LazyFrame>$rolling(
  index_column,
  ...,
  period,
  offset = NULL,
  closed = "right",
  group_by = NULL,
  check_sorted = TRUE
)

Arguments

`index_column`	Column used to group based on the time window. Often of type Date/Datetime. This column must be sorted in ascending order (or, if `by` is specified, then it must be sorted in ascending order within each group). In case of a rolling group by on indices, dtype needs to be either Int32 or Int64. Note that Int32 gets temporarily cast to Int64, so if performance matters use an Int64 column.
`…`	Ignored.
`period`	A character representing the length of the window, must be non-negative. See the `Polars duration string language` section for details.
`offset`	A character representing the offset of the window, or `NULL` (default). If `NULL`, `-period` is used. See the `Polars duration string language` section for details.
`closed`	Define which sides of the temporal interval are closed (inclusive). This can be either `“left”`, `“right”`, `“both”` or `“none”`.
`group_by`	Also group by this column/these columns.
`check_sorted`	Check whether data is actually sorted. Checking it is expensive so if you are sure the data within the `index_column` is sorted, you can set this to `FALSE` but note that if the data actually is unsorted, it will lead to incorrect output.

Details

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

"1i" \# length 1
"10i" \# length 10

Value

A LazyGroupBy object

Polars duration string language

Polars duration string language is a simple representation of durations. It is used in many Polars functions that accept durations.

It has the following format:

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)

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".

Examples

library(polars)

dates = c(
  "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$LazyFrame(dt = dates, a = c(3, 7, 5, 9, 2, 1))$with_columns(
  pl$col("dt")$str$strptime(pl$Datetime())$set_sorted()
)

df$rolling(index_column = "dt", period = "2d")$agg(
  sum_a = pl$sum("a"),
  min_a = pl$min("a"),
  max_a = pl$max("a")
)$collect()

#> shape: (6, 4)
#> ┌─────────────────────┬───────┬───────┬───────┐
#> │ dt                  ┆ sum_a ┆ min_a ┆ max_a │
#> │ ---                 ┆ ---   ┆ ---   ┆ ---   │
#> │ datetime[μs]        ┆ f64   ┆ f64   ┆ f64   │
#> ╞═════════════════════╪═══════╪═══════╪═══════╡
#> │ 2020-01-01 13:45:48 ┆ 3.0   ┆ 3.0   ┆ 3.0   │
#> │ 2020-01-01 16:42:13 ┆ 10.0  ┆ 3.0   ┆ 7.0   │
#> │ 2020-01-01 16:45:09 ┆ 15.0  ┆ 3.0   ┆ 7.0   │
#> │ 2020-01-02 18:12:48 ┆ 24.0  ┆ 3.0   ┆ 9.0   │
#> │ 2020-01-03 19:45:32 ┆ 11.0  ┆ 2.0   ┆ 9.0   │
#> │ 2020-01-08 23:16:43 ┆ 1.0   ┆ 1.0   ┆ 1.0   │
#> └─────────────────────┴───────┴───────┴───────┘