Dense Input
| Column | Data Type | Description |
|---|---|---|
| PartitionColumns | Same as InputTable | Columns on which input is partitioned. This appears only when the PartitionColumns argument is specified. |
| TD_STATTYPE_SCLFIT | VARCHAR (CHARACTER SET LATIN) |
Statistic names and parameters—see following table. |
| TargetColumns | REAL | Statistics values for TargetColumns argument. |
Sparse Input
| Column | Data Type | Description |
|---|---|---|
| PartitionColumns | Same as InputTable | Columns on which input is partitioned. This appears only when the PartitionColumns argument is specified. |
| TD_STATTYPE_SCLFIT | VARCHAR CHARACTER SET LATIN | Statistic names and parameters. |
| AttributeNameColumn | VARCHAR CHARACTER SET UNICODE | Name of the attributes. |
| AttributeValueColumn | REAL | Statistics values for the attributes. |
Stattypes the function outputs are:
| Statistic Name | Description |
|---|---|
| min | Minimum value in the corresponding column. |
| max | Maximum value in the corresponding column. |
| sum | Sum value in the corresponding column. |
| count | Count of valid values in the corresponding column. |
| null | Count of NULL values in the corresponding column. |
| avg | Average of valid values in corresponding column. |
| variance | Variance of values in corresponding column. Variance is calculated according to N-1 degrees of freedom. (number of valid values minus one). |
| ustd | Unbiased Standard deviation of values in corresponding column. Standard deviation is calculated according to N-1 degrees of freedom (number of valid values minus one). |
| std | Standard deviation of values in corresponding column. Standard deviation is calculated according to N degrees of freedom (number of valid values). |
| missvalue_* | * is the value of the MissValue argument: KEEP, ZERO, or LOCATION. |
| globalscale_* | * is the value of the GlobalScale argument : true or false. |
| unusedattributes_* | * is the value of the UnusedAttributes argument: unscaled or nullify. |
The values for location and scale parameters for different ScaleMethods are:
| Method | Description | Location | Scale |
|---|---|---|---|
| mean | Mean | Xmean | 1 |
| sum | Sum | 0 | |
| ustd | Unbiased Standard Deviation (Z-Score using Unbiased Standard Deviation) | Xmean | Standard deviation, calculated according to the unbiased estimator of the variance. ustd = sqrt ( sum( Xi - Xmean)^2 ) / N - 1) where N is the count of valid values |
| std | Standard Deviation (Z-Score using Standard Deviation) | Xmean | Standard deviation, calculated according to the biased estimator of the variance. std = sqrt ( sum( Xi - Xmean)^2 ) / N ) where N is the count of valid values |
| range | Range | Xmin | Xmax-Xmin |
| midrange | Midrange | (Xmax+Xmin)/2 | (Xmax+Xmin)/2 |
| maxabs | Maximum Absolute Value | 0 | Maximum of the absolute value of X. |
| rescale | Rescale using specified lower bound, upper bound, or both. | Lower_bound only: (Xmin - lb*) Upper_bound only: (Xmax - ub*) Lower_bound and Upper_bound: Xmin - (lb / (ub-lb) ) *where
|
Lower_bound only: 1 Upper_bound only: 1 Lower_bound and Upper_bound: (Xmax - Xmin) / (ub-lb) ) |