The function accepts only two ON clauses for input tables, where the Test table uses the PARTITION BY ANY clause, and the training table uses the dimension clause. Both the test and training tables are required.
TrainingTable Schema
| Column | Data Type | Description |
|---|---|---|
| IDColumn | BYTEINT,SMALLINT,INTEGER,BIGINT | The unique identifier of the training table. |
| InputColumns | NUMERIC | The columns that the function uses to compute the distance between test data points and the training data points. The test table must have corresponding columns in the training table with name and datatype. |
| ResponseColumn | NUMERIC (regression), INTEGER (classification) | The numeric value of the target variable used for prediction in KNN-based regression or classification. |
TestTable Schema
| Column | Data Type | Description |
|---|---|---|
| IDColumn | BYTEINT,SMALLINT,INTEGER,BIGINT | The unique identifier of the test data object. |
| InputColumns | NUMERIC | The columns that the function uses to compute the distance between test data points and the training data points. The training table must have corresponding columns in the test table with name and datatype. |