HeatWave Release Notes
MySQL 9.0.0 introduces the
ML_MODEL_ACTIVE routine. Use
this routine to check which models are active for which users.
All active users and models share the amount of memory defined
by the shape, and it might be necessary to schedule users.
As of MySQL 9.0.0,
ML_MODEL_ACTIVE can report
which models are active for which users.
mysql> CALL sys.ML_MODEL_ACTIVE (user, model_info);
ML_MODEL_ACTIVE parameters:
user: The user to provide information for. Set tocurrentorallorNULL.NULLis equivalent tocurrent.-
model_info: The name of the JSON array that will contain the active user and model information. It will contain two JSON object literals:-
A JSON object literal that will contain the following :
Key:
total model size(bytes)-
Value:
If
user: is set tocurrent: The sum of model sizes for the current user.If
user: is set toall: The sum of model sizes for all users.
-
A JSON object literal that will contain the following :
-
Key:
If
user: is set tocurrent: The model handle for an active model owned by the current user.If
user: is set toall: The name of a user who has active models.
-
Value:
If
user: is set tocurrent: Themodel_metadatafor this model.-
If
user: is set toall: A list of JSON object literals:Key: The model handle.
Value: Brief model metadata for that model.
-
-
-
user1checks their own models:mysql> CALL sys.ML_MODEL_ACTIVE('current', @model_info); Query OK, 0 rows affected (0.10 sec) mysql> SELECT JSON_PRETTY(@model_info); +-----------------------------------------------------------+ | JSON_PRETTY(@model_info) | +-----------------------------------------------------------+ | [ | { | "total model size(bytes)": 348954 | }, | { | "iris_export_user1": { | "task": "classification", | "notes": "", | "chunks": 1, | "format": "HWMLv2.0", | "n_rows": 120, | "status": "Ready", | "options": { | "model_explainer": "permutation_importance, shap", | "prediction_explainer": "shap" | }, | "n_columns": 4, | "pos_class": null, | "column_names": [ | "sepal length", | "sepal width", | "petal length", | "petal width" | ], | "contamination": null, | "model_quality": "high", | "training_time": 18.363686, | "algorithm_name": "ExtraTreesClassifier", | "training_score": -0.10970368035588404, | "build_timestamp": 1697524180, | "n_selected_rows": 96, | "training_params": { | "sp_arr": null, | "timezone": null, | "recommend": "ratings", | "force_use_X": false, | "recommend_k": 3, | "remove_seen": true, | "contamination": null, | "feedback_threshold": 1 | }, | "train_table_name": "mlcorpus.iris_train", | "model_explanation": { | "shap": { | "petal width": 0.3139, | "sepal width": 0.0296, | "petal length": 0.2787, | "sepal length": 0.0462 | }, | "permutation_importance": { | "petal width": 0.2301, | "sepal width": 0.0056, | "petal length": 0.2192, | "sepal length": 0.0056 | } | }, | "model_object_size": 348954, | "n_selected_columns": 4, | "target_column_name": "class", | "optimization_metric": "neg_log_loss", | "selected_column_names": [ | "petal length", | "petal width", | "sepal length", | "sepal width" | ] | } | } | ] | | +-----------------------------------------------------------+ 1 row in set (0.00 sec) -
user1checks their own models, and extracts specific information:mysql> CALL sys.ML_MODEL_ACTIVE('current', @model_info); Query OK, 0 rows affected (0.12 sec) mysql> SELECT JSON_KEYS(JSON_EXTRACT(@model_info, '$[1]')); +------------------------------------------------------------+ | JSON_KEYS(JSON_EXTRACT(@model_info, '$[1]')) | +------------------------------------------------------------+ | ["iris_export", "mlcorpus.iris_train_user1_1697524152037"] | +------------------------------------------------------------+ 1 row in set (0.00 sec) mysql> SELECT JSON_EXTRACT(@model_info, '$[0]'); +--------------------------------------+ | JSON_EXTRACT(@model_info, '$[0]') | +--------------------------------------+ | {"total model size(bytes)": 697908} | +--------------------------------------+ 1 row in set (0.01 sec) -
user1checks the models for all users:mysql> CALL sys.ML_MODEL_ACTIVE('all', @model_info); Query OK, 0 rows affected (0.11 sec) mysql> SELECT JSON_PRETTY(@model_info); +-----------------------------------------------------+ | JSON_PRETTY(@model_info) | +-----------------------------------------------------+ | [ | { | "total model size(bytes)": 1046862 | }, | { | "user2": [ | { | "iris_export_user2": { | "format": "HWMLv2.0", | "model_size(byte)": 348954 | } | } | ], | "user1": [ | { | "mlcorpus.iris_train_user1_1697524152037": { | "format": "HWMLv2.0", | "model_size(byte)": 348954 | } | }, | { | "iris_export": { | "format": "HWMLv2.0", | "model_size(byte)": 348954 | } | } | ] | } | ] | | +-----------------------------------------------------+ 1 row in set (0.00 sec) -
user2checks the models for all users:mysql> CALL sys.ML_MODEL_ACTIVE('all', @model_info); Query OK, 0 rows affected (0.10 sec) mysql> SELECT JSON_PRETTY(@model_info) +-----------------------------------------------------+ | JSON_PRETTY(@model_info) | +-----------------------------------------------------+ | [ | { | "total model size(bytes)": 1046862 | }, | { | "user2": [ | { | "iris_export_user2": { | "format": "HWMLv2.0", | "model_size(byte)": 348954 | } | } | ], | "user1": [ | { | "mlcorpus.iris_train_user1_1697524152037": { | "format": "HWMLv2.0", | "model_size(byte)": 348954 | } | }, | { | "iris_export": { | "format": "HWMLv2.0", | "model_size(byte)": 348954 | } | } | ] | } | ] | | +-----------------------------------------------------+ 1 row in set (0.01 sec)