0

I am using featuretools 0.20.0 and koalas 1.3.0.

create feature matrix for all customers

feature_matrix_cust, feature_defs = ft.dfs( enter code here`entityset=es4, target_entity="customers_ks", agg_primitives=["count", "avg_time_between", "num_unique", "trend"], where_primitives=["count", "avg_time_between","num_unique", "trend"], trans_primitives=["time_since_previous"] )

I got error below: ValueError: Selected primitives are incompatible with Koalas EntitySets: time_since_previous, avg_time_between, trend, avg_time_between, trend

Will featuretools support those primitives for Koalas in the near future? Is there anyway to handle those primitives with current featuretools and Koalas version?

hvvo2004
  • 1

1 Answers1

1

You may be able to write custom primitives that include Koalas functionality, although it is worthwhile to note that currently Koalas does not support custom groupby aggregation functions. Improved support for distributed entitysets is ongoing, however there's currently no timeline for when support for additional primitives will be added.

Frances Hartwell
  • 191
  • 1