Skip to contents

Resample a time series to a calendar grid

Source: R/resample.R
resample.Rd

Buckets a datetime column to a grid (every) and aggregates within each bucket – the time-series form of group_by() + summarise(). Equivalent to mutate(<time> = floor_time(<time>, every)), group_by(<time>), then summarise(...).

Usage

resample(.data, time, every, ..., .name = NULL)

Arguments

.data

A vectra_node.

time

The datetime column to bucket (unquoted).

every

Bucket size as a string, e.g. "1 hour", "15 min", "day", "month" (see floor_time()).

...

Named aggregation expressions, as in summarise() (e.g. mean_temp = mean(temp)).

.name

Optional name for the bucket column. Defaults to the name of time (the original column, replaced by its floored value).

Value

A vectra_node with one row per occupied bucket: the bucket column followed by the aggregates. The bucket collects as a numeric epoch value (see floor_time() on restoring the date class).

See also

floor_time(), summarise()

Examples

if (FALSE) { # \dontrun{
tbl("readings.vtr") |>
  resample(timestamp, "1 hour", mean_temp = mean(temp), n = n()) |>
  collect()
} # }