This project is now called Atoti.
These are the technical changes between
atoti and the last published version of
Harmonization of the aggregation API. Aggregation operations include store field aggregation such as
atoti.agg.sumbut also other more “advanced” operations such as
running_avg, etc. In
atoti, all aggregation operations involving sums are performed with atoti.agg.sum, all operations dealing with the computation of a minimum are performed with atoti.agg.min, and so on. “Advanced” measures like cumulative sum measures, dynamic aggregation measures, or siblings aggregation measures, can be defined using the new
scopeargument in every aggregation function. The scope can be created with atoti.scope, and is one of:
atoti.scope.cumulative() to compute the aggregated value of some other members in the same hierarchy
atoti.scope.origin() to specify one or more levels above which the values are computed using the new aggregation function
atoti.scope.siblings() to aggregate the values of all siblings in a hierarchy
New behaviour of store.join(). All hierarchies created during a join are non-slicing. The
slicing_columnsargument is thus gone. If you want the created hierarchies to be slicing, you can set the Hierarchy.slicing property to
We improved our Python library and JupyterLab extension Conda packages to ensure you can’t install incompatible version of each other in your environment. We thus stopped providing a “bundle” package. You can follow the new installation procedure.