Skip to contents

Group a dataset's time slices into calendar periods.

group_by_time.Rd

Partitions every band's slices by period so a following reduce_over(over = "t") builds one composite per group: a year of daily imagery, group_by_time("month"), then a median gives twelve monthly composites (xarray's resample(time = ...).reduce()). Slices are grouped by the period prefix of their date name, so build the dataset at granularity = "day" and group up from there. Ragged bands are fine – a band with no slice in a group is simply absent from that group's composite.

Usage

group_by_time(x, by = "month")

Arguments

x

A LazyDataset.

by

"year", "quarter", "month" (default), "week", "day", or a function mapping a slice name to a group label.

Value

A LazyDatasetGroups (a named list of per-group LazyDatasets). Reduce it with reduce_over(), then collect() returns a named list of results (or writes one file per group when path carries a {group} placeholder, e.g. "ndvi_{group}.tif").