The OutputTable schema depends on the OutputFormat syntax element.
OutputTable Schema, OutputFormat ('sparse') (Default)
The table has D+3 columns, where D is the number of dimensions of the input data.
| Column | Data Type | Description |
|---|---|---|
| accumulate_column | Same as in input table | [Column appears once for each specified accumulate_column.] Column copied from InputTable. Typically, one accumulate_column contains unique data point identifier. |
| data_point | Any numeric SQL data type | [Column appears D times.] Input data point, copied from InputTable. |
| cluster_rank | INTEGER | Cluster rank, from most probable to least probable. Most probable cluster is identified as "1", next most probable as "2", and so on. |
| cluster_id | INTEGER | Cluster identification number. |
| prob | DOUBLE PRECISION | Probability that cluster is assigned to data point. |
OutputTable Schema, OutputFormat ('dense')
The table has D+2k columns, where D is the number of dimensions of the input data and n is the the number of cluster weights that the function outputs (the value of the TopK syntax element).
| Column | Data Type | Description |
|---|---|---|
| accumulate_column | Same as in input table | [Column appears once for each specified accumulate_column.] Column copied from InputTable. Typically, one accumulate_column contains unique data point identifier. |
| id | Any | Data point identifier. |
| data_point | Any numeric SQL data type | [Column appears D times.] Input data point, copied from InputTable. |
| cluster_id_i | INTEGER | [Column appears n times.] cluster_id_1 is most probable cluster for observation, cluster_id_2 is next most probable, and so on. |
| prob_i | DOUBLE PRECISION | [Column appears n times.] Probability that observation belongs to cluster_id_i. |