The Forest_Predict function has a required input table and an optional model table.
If you do not specify the optional model table, then you must specify the model with either the ModelFile or Forest argument. Teradata recommends that you specify the optional model table.
The input table for the Forest_Predict function must contain an ID column (for example, user_id or transaction_id), so that each test point can be associated with a prediction. It must also contain all columns specified by the NumericInputs argument and all columns specified by the CategoricalInputs argument. The following table shows its schema.
| Column Name | Data Type | Description |
|---|---|---|
| id_column | Any | Contains unique identifiers for test points in the test set. Cannot contain NULL values. |
| numeric_column | NUMERIC, INTEGER, BIGINT, or DOUBLE PRECISION | Contains numeric inputs. Cannot contain NULL values. |
| category_column | INTEGER, BIGINT, or VARCHAR | Contains categorical inputs. Cannot contain NULL values. |
| accumulate_column | Any | Column to copy to the output table. |