Type Members

1.  class ColToTensor extends RowTransformSchema

   A schema which specialize on transforming single column to size(1) Tensor.

   A schema which specialize on transforming single column to size(1) Tensor. Types of selected columns will be identified according to their DataType. And type conversions will be done automatically from DataType to TensorDataType.

2.  class ColsToNumeric[T] extends RowTransformSchema

   A schema which specialize on transforming multiple numeric columns to one Tensor.

   A schema which specialize on transforming multiple numeric columns to one Tensor. Types of selected columns will be identified according to their [DataType. And type conversions will be done automatically from DataType to T if valid. Currently, DoubleType, FloatType, ShortType, IntegerType, LongType are supported.

   T

   the type of result Tensor

3.  trait RowTransformSchema extends Serializable

   A schema describe a transforming job which convert a Row to a Table(Tensor).

4.  class RowTransformer extends Transformer[Row, Table]

   This transformer is a container of RowTransformSchemas.

   RowTransformer transform a Row to a Table whose values are all Tensor.

   This transformer is a container of RowTransformSchemas. When this transformer being executed, it will run transform methods of its RowTransformSchemas.

   Output of RowTransformer is a Table. The keys of Table are Tensor.scalar(schemaKey)s of included RowTransformSchemas. Correspondingly, the values of Table are results(Tensor) of RowTransformSchema.transform.