Problem
DataStore.groupby(...) returns LazyGroupBy, but two pandas-standard iteration patterns don't work:
import pandas as pd
from datastore import DataStore
pdf = pd.DataFrame({
'date': ['2026-05-22', '2026-05-22', '2026-05-23'],
'code': ['000001', '000002', '000001'],
'price': [10.0, 20.0, 10.5],
})
ds = DataStore(pdf)
# 1) DataFrameGroupBy iteration
for (date, code), group in ds.groupby(['date', 'code']):
...
# TypeError: Expected str or list, got int
# (misleading - __iter__ is missing, Python falls back to __getitem__(0))
# 2) SeriesGroupBy iteration
for k, s in ds.groupby('date')['price']:
print(k, s)
# Yields raw scalar values, not (key, sub_series) pairs - silently wrongpandas equivalents work fine and are very common (per-group loops, plotting, custom logic). The error in case 1 is especially misleading: it hides a missing feature behind a wrong-arg-type message.
Expected
Mirror pandas semantics for LazyGroupBy (DataFrameGroupBy) and the ColumnExpr returned by gb['col'] (SeriesGroupBy): iteration, get_group(), .groups, .indices, len(), in.
Fix
PR with implementation + 44 mirror-pattern tests incoming.
Problem
DataStore.groupby(...)returnsLazyGroupBy, but two pandas-standard iteration patterns don't work:pandas equivalents work fine and are very common (per-group loops, plotting, custom logic). The error in case 1 is especially misleading: it hides a missing feature behind a wrong-arg-type message.
Expected
Mirror pandas semantics for
LazyGroupBy(DataFrameGroupBy) and theColumnExprreturned bygb['col'](SeriesGroupBy): iteration,get_group(),.groups,.indices,len(),in.Fix
PR with implementation + 44 mirror-pattern tests incoming.