Dear Exasol,
the documentation provides an example of an analytic user-defined aggregate function my_sum: UDF Scripts - Examples | Exasol DB Documentation
This function can be executed using the following syntax:
SELECT my_sum(x ORDER BY x) FROM t;
Here, the ORDER BY clause acts similarly to the common SQL window function pattern:
aggregate(x) OVER (ORDER BY some_column)
Would it be possible to extend this feature to also support the PARTITION BY clause in UDFs?
In other words, an equivalent to the syntax:
aggregate(x) OVER (PARTITION BY some_column)
Such an extension would be very useful for more granular, partition-aware UDFs, for example in the implementation of the ML clustering.
Looking forward to your feedback!