model_name!PREDICT¶

Syntax¶

<model_name>!PREDICT(
    INPUT_DATA => <input_data>,
    [CONFIG_OBJECT => <config_object>]
)

Arguments¶

Required

INPUT_DATA

An OBJECT containing key-value pairs that map feature names to their values. Use wildcard expansion in an object literal to automatically create key-value pairs from a table, as in:

SELECT model_binary!PREDICT(INPUT_DATA => {*})
    as prediction from prediction_purchase_data;

Copy

The feature names in the object must match the names and types specified at training time. Missing or extraneous features are ignored.

Optional

CONFIG_OBJECT

An OBJECT whose key-value pairs specify additional training options.

Key	Type	Default	Description
on_error	STRING	'ABORT'	String (constant) that specifies the error handling method for the model inference task. Supported values are: 'ABORT': Abort the entire prediction operation if any row results in an error. 'SKIP': Skip rows that result in an error. The error is shown instead of the results.

Output¶

Column	Type	Description
PREDICTION	VARIANT	Prediction results as an OBJECT containing the following keys. Key Type Description class STRING The predicted label with the highest probability. probability VARIANT An OBJECT containing the probabilities of each predicted class. For each class, the key is the class name, and the value is the predicted probability of the class.
LOGS	VARIANT	Contains error or warning messages.

Examples¶

See Examples.