datacube.api.Tile

class datacube.api.Tile(sources, geobox)[source]

The Tile object holds a lightweight representation of a datacube result.

It is produced by GridWorkflow.list_cells() or GridWorkflow.list_tiles().

The Tile object can be passed to GridWorkflow.load() to be loaded into memory as an xarray.Dataset.

A portion of a tile can be created by using index notation. eg:

tile[0:1, 0:1000, 0:1000]

This can be used to load small portions of data into memory, instead of having to access the entire Tile at once.

__init__(sources, geobox)[source]

Create a Tile representing a dataset that can be loaded.

Parameters:
  • sources (xarray.DataArray) – An array of non-spatial dimensions of the request, holding lists of datacube.storage.DatasetSource objects.
  • geobox (model.GeoBox) – The spatial footprint of the Tile

Methods

__init__(sources, geobox) Create a Tile representing a dataset that can be loaded.
split(dim[, step]) Splits along a non-spatial dimension into Tile objects with a length of 1 or more in the dim dimension.
split_by_time([freq, time_dim]) Splits along the time dimension, into periods, using pandas offsets, such as:

Attributes

dims Names of the dimensions, eg ('time', 'y', 'x') :return: tuple(str)
product datacube.model.DatasetType
shape Lengths of each dimension, eg (285, 4000, 4000) :return: tuple(int)