id_column |
Any |
Specified column name that uniquely identifies an observation in test table. |
partition_by_column |
CHARACTER, VARCHAR, INTEGER, BIGINT, SMALLINT, BYTEINT |
[Partition by key only] Unique identifier of model in InputTable. |
prediction |
FLOAT |
Predicted value of the test observation. |
prob |
FLOAT |
Probability that the observation belongs to the predicted class. Only appears if the OutputProb element is set to True and the Responses element is not specified. |
prob_0 |
FLOAT |
Probability that the observation belongs to class 0. Only appears if the Responses element is specified. |
prob_1 |
FLOAT |
The probability that the observation belongs to class 1. Only appears if the Responses element is specified. |
accumulate_column |
Any |
Specified column names in the Accumulate element copied to the output table. |
error_code |
VARCHAR |
Error code shown if the function cannot predict. The codes are as follows:- 1: Broken model. The training function fails to generate a model.
- 2: Missing model. The prediction input table has partition_by_column that does not appear in the training model table.
- 3: Invalid row. During prediction, the input table has invalid rows. For example, a row may contain NULL value that cannot be processed.
- 4: Invalid column type. During prediction, the input table has invalid column type. For example, a column may need to be numeric type instead of character varying type.
- -1: Unknown.
|