You can apply a window specification to an aggregate function. The following window types are supported for aggregate UDFs:
| Window Type | Aggregation Group | Partitioning Strategy |
|---|---|---|
| Reporting window | ROWS BETWEEN UNBOUNDED PRECEDING AND UNBOUNDED FOLLOWING | Hash partitioning |
| Cumulative window | ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW or ROWS BETWEEN CURRENT ROW AND UNBOUNDED FOLLOWING |
Hash partitioning |
| Moving window |
|
Hash partitioning and value partitioning |
The following window types are not supported for aggregate UDFs:
| Window Type | Aggregation Group |
|---|---|
| Moving window |
|
The partitioning strategy helps to avoid hot AMP situations where the values of the columns of the PARTITION BY clause result in the distribution of too many rows to the same partition or AMP. You should make sure that the method uses the right set of columns for the PARTITION BY clause to avoid potential skew situations for the reporting or cumulative aggregate cases. For more information, see “Data in Partitioning Column of Window Specification and Resource Impact” in Teradata Vantage™ - SQL Functions, Expressions, and Predicates, B035-1145.