array Driver

The array driver specifies a TensorStore backed by a contiguous in-memory array.

When specified via JSON, this driver is useful for specifying small, constant arrays.

When used from the C++ or Python API, this driver adapts an existing in-memory array for use as a TensorStore.

json driver/array : object
  • TensorStore — Specifies a TensorStore to open/create.

Required members:
driver : "array"
dtype : dtype

Specifies the data type.


Nested array of rank equal to rank specifying the initial value.

Optional members:
context : Context

Specifies context resources that augment/override the parent context.

rank : integer[0, 32]

Specifies the rank of the TensorStore.

If transform is also specified, the input rank must match. Otherwise, the rank constraint applies to the driver directly.

transform : IndexTransform

Specifies a transform.

schema : Schema

Specifies constraints on the schema.

When opening an existing array, specifies constraints on the existing schema; opening will fail if the constraints do not match. Any soft constraints specified in the chunk_layout are ignored. When creating a new array, a suitable schema will be selected automatically based on the specified schema constraints in combination with any driver-specific constraints.

data_copy_concurrency : ContextResource = "data_copy_concurrency"

Specifies or references a previously defined Context.data_copy_concurrency. It is normally more convenient to specify a default data_copy_concurrency in the context.


{"driver": "array", "array": [[1, 2, 3], [4, 5, 6]], "dtype": "int32"}