pyflink.dataframe.DataType.tensor#

classmethod DataType.tensor(dtype: pyflink.dataframe.datatype.DataType, shape: tuple) → pyflink.dataframe.datatype.DataType[source]#

Fixed-shape tensor type.

Uses the first-class TENSOR logical type so that shape metadata is carried natively by Flink’s type system rather than via a sideband metadata mechanism.

Parameters

dtype – Element type, e.g. DataType.float32() or DataType.int8().
shape – Shape tuple, e.g. (3, 224, 224). Must be fully specified (no dynamic dimensions).

Example:

>>> DataType.tensor(DataType.float32(), shape=(3, 224, 224))
>>> DataType.tensor(DataType.int8(),    shape=(224, 224, 3))