This PR adds aggregation function support for Grizzly Series. The API follows that of Pandas, with the exception of the index of the returned value: Pandas assigns labels based on the aggregation function, while Grizzly (which does not support custom indexes at the moment) assigns numerical labels.
Multiple aggregations are co-optimized to prevent redundant computation, e.g., a requesting a variance and a mean will only compute the mean once and use the result for variance computation.
This PR adds aggregation function support for Grizzly Series. The API follows that of Pandas, with the exception of the index of the returned value: Pandas assigns labels based on the aggregation function, while Grizzly (which does not support custom indexes at the moment) assigns numerical labels.
Examples of aggregations with Grizzly
Multiple aggregations are co-optimized to prevent redundant computation, e.g., a requesting a variance and a mean will only compute the mean once and use the result for variance computation.
Currently supports:
Some of the IR code is taken from https://www.github.com/weld-project/baloo.