Argument | Category | Description |
---|---|---|
ModelType | Required | Specifies the function name for which this .aml file is to be used. For predictors, this represents the type of the model trained. The model type of functions are listed in Aster Scoring SDK Functions. |
ModelTable | Optional | Specifies the input model table representing input data for which aml file is to be generated. The argument clause accepts multiple input model tables. These tables could be either database tables or files installed on the database. |
ModelTag | Optional | Specifies the input model table tag for each input model table specified in ModelTable clause. Tags are supported for functions which accept multiple model tables and are used to distinguish the role of each model table in the context of the Aster Scoring SDK function. The supported tags are listed in Aster Scoring SDK Functions. The number of entries must match the number of entries in the ModelTable clause. |
InstalledFile | Optional | Specifies whether the corresponding value in ModelTable is a file installed on database or a table on database. If ModelTable is a database table, set this to false (default) and if it is a file, set this to true. The number of entries must match the number of entries in the ModelTable clause. |
RequestColNames | Required | Specifies the column names of the request to be scored. For a predictor, these column names typically match the columns names of the training data used to train the model. |
RequestColTypes | Required | Specifies the column types of the request to be scored. For a predictor, these column types typically match the columns types of the training data used to train the model. |
AMLPrefix | Optional | Specifies the name of the generated AML model file. Default value is “model”. The output file is stored with suffix .aml. |
OverwriteOutput | Optional | Specifies whether the output AML model file is to be overwritten if it already exists. |
RequestArgNamen | Optional | Specifies the argument clause name for the function to be used in scoring. These clauses are the same as mentioned in the corresponding SQL-MapReduce function documentation, with some exceptions. The supported clauses for each function are listed in Aster Scoring SDK Functions. The RequestArgName must start from 1 in a sequential manner. For example, the following sequence of clauses is acceptable: RequestArgName1, RequestArgName2, RequestArgName3. However, this sequence is not acceptable: RequestArgName1, RequestArgName2, RequestArgName4. The maximum value of n is 30. The values for corresponding arguments are provided in RequestArgVal clause. |
RequestArgVal[n] | Optional | Specifies the argument clause value for the function to be used in scoring. The value in this clause corresponds to the clause mentioned in RequestArgName for the same value of n. These clause values are the same as mentioned in the documentation for the corresponding SQL-MapReduce function. The RequestArgVal must start from 1 in a sequential manner, and follows the same format as RequestArgName. The maximum value of n is 30. |