Skip to content
Chelis docs

GroupBy

Coral.GroupBy currently validates a single-key aggregation slice against checked-in pandas goldens. The current implementation uses host-path key equality and preserves first-seen key order (sort=False-style behavior).

Validated operations in the current slice:

  • group_by
  • agg_sum
  • agg_mean
  • agg_count
  • agg_min
  • agg_max
  • agg for explicit multi-aggregation specs in requested order
  • value_counts: frequency table for a single column (returns a Frame with the key column and a "count" column)
module Coral.BookGroupBy
import Coral.Frame (from_pairs, nrows, int_col_of_list)
import Coral.GroupBy (group_by, agg)
export (main)
def main() -> int64 = {
  frame = from_pairs([("city", StringCol(["london", "paris", "london"])), ("qty", int_col_of_list([cast(5, int64), cast(6, int64), cast(7, int64)])), ("price", FloatCol(to_tensor([cast(10.0, f32), cast(20.0, f32), cast(30.0, f32)])))])
  grouped = group_by(frame, "city")
  totals = agg(grouped, [("qty", AggSum), ("price", AggMean)])
  nrows(totals)
}

The gate today is fixture-backed plus compile-checked. Stripped bare builds for GroupBy are clean on chelis v0.7.6: build, link, and execution all pass with no upstream blockers.