MySQL :: MySQL HeatWave User Guide :: 3.12 HeatWave ML Error Messages

Error number: ML001016; SQLSTATE: HY000

Message: Only classification and regression tasks are supported

Check the task option in your ML_TRAIN call to ensure that it was specified correctly.

Error number: ML001031; SQLSTATE: HY000

Message: samples per class, and cannot be trained on. For a real valued target column, the task parameter in the options JSON should be set to regression.

Example: ERROR HY000: "ML001031: Running as a classification task. 189 classes have less than 5 samples per class, and cannot be trained on. Maybe it should be trained as a regression task instead of a classification task. Or the task ran on the default setting - classification, due to an incorrect JSON task argument"

If a classification model is intended, add more samples to the data to increase the minority class count; that is, add more rows with the underrepresented target column value. If a classification model was not intended, run ML_TRAIN with the regression task option.

Error number: ML001051; SQLSTATE: HY000

Message: One or more rows contain all NaN values

Example: ERROR HY000: "ML001051: One or more rows contain all NaN values. Imputation is not possible on such rows."

MySQL does not support NaN values. Replace with NULL.

Error number: ML001052; SQLSTATE: HY000

Message: All columns are dropped. They are constant, mostly unique, or have a lot of missing values!

Example: ERROR HY000: "ML001052: All columns are dropped. They are constant, mostly unique, or have a lot of missing values!"

ML_TRAIN ignores columns with certain characteristics such as columns missing more than 20% of values and columns containing the same single value. See Section 3.3, “Preparing Data”.

Error number: ML001053; SQLSTATE: HY000

Message: Unlabeled samples detected in the training data.

Example: ERROR HY000: "ML001053: Unlabeled samples detected in the training data. (Values in target column can not be NULL)"

Training data must be labeled. See Section 3.3, “Preparing Data”.

Error number: ML003000; SQLSTATE: HY000

Message: Number of offloaded datasets has reached the limit!

Error number: ML003011; SQLSTATE: HY000

Message: Columns of provided data need to match those used for training

Example: ERROR HY000: "ML003011: Columns of provided data need to match those used for training. Provided - ['petal length', 'petal width', 'sepal length', 'sepal width'] vs Trained - ['petal length', 'sepal length', 'sepal width']"

Possible causes: The input data columns do not match the columns of training dataset used to train the model. Compare your input data to your training data to identify the discrepancy.

Error number: ML003012; SQLSTATE: HY000

Message: is NULL or has not been loaded

Example: ERROR HY000: "ML003012: 'The table (mlcorpus.iris_train) is NULL or has not been loaded.'"

Possible causes: There is no data in the specified table.

Error number: ML003014; SQLSTATE: HY000

Message: The size of model generated is larger than the maximum allowed

Possible causes: Models greater than 900 MB in size are not supported.

Error number: ML003015; SQLSTATE: HY000

Message: The input column types do not match the column types of dataset which the model was trained on

Example: ERROR HY000: "ML003015: The input column types do not match the column types of dataset which the model was trained on. ['numerical', 'numerical', 'categorical', 'numerical'] vs ['numerical', 'numerical', 'numerical', 'numerical']"

Error number: ML003016; SQLSTATE: HY000

Message: in input JSON ->

Example: ERROR HY000: "ML003016: Missing argument \"row_json\" in input JSON -> dict_keys(['operation', 'user_name', 'table_name', 'schema_name', 'model_handle'])"

Possible causes: The syntax of your ML routine call is not valid.

Error number: ML003017; SQLSTATE: HY000

Message: The corresponding value of row_json should be a string!

Example: ERROR HY000: "ML003017: The corresponding value of row_json should be a string!"

Possible causes: The syntax of your ML routine call is not valid.

Error number: ML003018; SQLSTATE: HY000

Message: The corresponding value of row_json is NOT a valid JSON!

Example: ERROR HY000: "ML003018: The corresponding value of row_json is NOT a valid JSON!"

Possible causes: The syntax of your ML routine call is not valid.

Error number: ML003019; SQLSTATE: HY000

Message: Invalid data for the metric

Example: ERROR HY000: "ML003019: Invalid data for the metric (roc_auc). Score could not be computed"

Possible causes: The scoring metric is legal and supported, but the data provided is not suitable to calculate such score; for example: ROC_AUC for multi-class classification. Try a different scoring metric.

Error number: ML003020; SQLSTATE: HY000

Message: Unsupported scoring function

Example: ERROR HY000: "ML003020: Unsupported scoring function (accuracy) for current task (regression)."

Possible causes: The scoring metric is legal and supported, but the task provided is not suitable to calculate such score; for example: Using the accuracy metric for a regression model.

Error number: ML003021; SQLSTATE: HY000

Message: Cannot train a regression task with a non-numeric target column.

Possible causes: ML_TRAIN was run with the regression task type on a training dataset with a non-numeric target column. Regression models require a numeric target column.

Error number: ML003022; SQLSTATE: HY000

Message: At least 2 target classes are needed for classification task.

Possible causes: ML_TRAIN was run with the classification task type on a training dataset where the target column did not have at least two possible values.

Error number: ML003023; SQLSTATE: HY000

Message: Unknown option given. Allowed options for training are.

Possible causes: The ML_TRAIN call specified an unknown option.

Error number: ML003024; SQLSTATE: HY000

Message: Not enough available memory, unloading any RAPID tables will help to free up memory.

Possible causes: There is not enough memory on your HeatWave Cluster to perform the operation. Try unloading data that was loaded for analytics to free up space.

For HeatWave Machine Learning functionality, the HeatWave.256GB node shape is recommended. The HeatWave.16GB node shape may not have enough memory for training on large data sets. If you see this error message when using the smaller node shape with HeatWave ML, use the larger shape instead.

Error number: ML004014; SQLSTATE: HY000

Message: Missing expected JSON key

Example: ERROR HY000: ML004014: Missing expected JSON key (schema_name).

Error number: ML004015 ; SQLSTATE: HY000

Message: Expected JSON string type value for key

Example: ERROR HY000: ML004015: Expected JSON string type value for key (schema_name).

Error number: ML004016; SQLSTATE: HY000

Message: Given JSON is larger than maximum permitted size

Example: ERROR HY000: ML004016: Given JSON (prediction_row prediction_row prediction_row prediction_row prediction_row prediction_row prediction_row prediction_row prediction_row) is larger than maximum permitted size

Error number: ML004018; SQLSTATE: HY000

Message: Parsing JSON arg: failed!

Example: ERROR HY000: ML004018: Parsing JSON arg: Invalid value. failed!

Error number: ML004019; SQLSTATE: HY000

Message: Expected JSON object type value for key

Example: ERROR HY000: ML004019: Expected JSON object type value for key (JSON root).

Error number: ML004020; SQLSTATE: HY000

Message: Operation was interrupted by the user

If a user-initiated interruption (Ctrl-C) is detected during the first phase of HeatWave ML model and table load where a MySQL parallel scan is used in the HeatWave plugin to read data from MySQL database and send it to the HeatWave Cluster, error messaging is handled by the MySQL parallel scan function and directed to ERROR 1317 (70100): Query execution was interrupted. The ERROR 1317 (70100) message is reported to the client instead of the ML004020 error message.

Error number: ML004022; SQLSTATE: HY000

Message: The user doesn't have access privileges to %.

Example: ERROR HY000: ML004022: The user doesn't have access privileges to ml.foo

Error number: ML004026; SQLSTATE: HY000

Message: A column (%) with an unsupported column type (%) detected!

Example: ERROR HY000: ML004026: A column (D1) with an unsupported column type (DATETIME) detected!

Error number: ML004051; SQLSTATE: HY000

Message: Invalid operation

Example: ERROR HY000: ML004051: Invalid operation

Error number: ML004999; SQLSTATE: HY000

Message: Error during Machine Learning