`statistics` on `NetworkCollection` not working reliably with different `snapshot_weightings`

[euronion]

Version Checks (indicate both or one)

• I have confirmed this bug exists on the lastest release of PyPSA.

• I have confirmed this bug exists on the current master branch of PyPSA.

Issue Description

Two networks loaded into the NetworkCollection with different snapshot_weightings:

Calling nc.statistics() seems to work fine:

However calling sub-statistics does not seem to reliably work. Calling nc.statistics.supply() only returns the supply-statistics for the hourly model:

Calling nc.statistics()["Supply"] however seems to work:

Reproducible Example

-

Expected Behavior

Installed Versions

Details

Replace this line.

[euronion]

Addendum: In the example above, the index.names are also messed up:

And what is worse, when accessing nc.statistics()["Supply"] the numbers are shown for the time-segmented networks, but they are wrong:

[tgilon]

I have a similar issue with a collection where one of the two networks has one day less (last day of the year). I'm unable to compute any time-related statistics (any function using _aggregate_timeseries).

Here is my Traceback:

Details

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
Cell In[161], line 1
----> 1 nc2030.statistics.opex()

File open-tyndp/.pixi/envs/open-tyndp/lib/python3.12/site-packages/pypsa/statistics/expressions.py:258, in StatisticHandler.__call__(self, *args, **kwargs)
    257 def __call__(self, *args: Any, **kwargs: Any) -> pd.DataFrame:  # noqa: D102
--> 258     return self._bound_method(*args, **kwargs)

File open-tyndp/.pixi/envs/open-tyndp/lib/python3.12/site-packages/pypsa/common.py:531, in deprecated_kwargs.<locals>.deco.<locals>.wrapper(*args, **kwargs)
    526 @functools.wraps(f)
    527 def wrapper(*args: Any, **kwargs: Any) -> Any:
    528     rename_deprecated_kwargs(
    529         f.__name__, kwargs, aliases, deprecated_in, removed_in
    530     )
--> 531     return f(*args, **kwargs)

File open-tyndp/.pixi/envs/open-tyndp/lib/python3.12/site-packages/pypsa/statistics/expressions.py:1612, in StatisticsAccessor.opex(self, components, groupby_time, groupby_method, aggregate_across_components, groupby, at_port, carrier, bus_carrier, nice_names, drop_zero, round, cost_types)
   1609     result = pd.concat(result)
   1610     return result.groupby(level=list(range(result.index.nlevels))).sum()
-> 1612 df = self._aggregate_components(
   1613     func,
   1614     components=components,
   1615     agg=groupby_method,
   1616     aggregate_across_components=aggregate_across_components,
   1617     groupby=groupby,
   1618     at_port=at_port,
   1619     carrier=carrier,
   1620     bus_carrier=bus_carrier,
   1621     nice_names=nice_names,
   1622     drop_zero=drop_zero,
   1623     round=round,
   1624 )
   1625 df.attrs["name"] = "Operational Expenditure"
   1626 df.attrs["unit"] = "currency"

File open-tyndp/.pixi/envs/open-tyndp/lib/python3.12/site-packages/pypsa/statistics/abstract.py:200, in AbstractStatisticsAccessor._aggregate_components(self, func, agg, components, groupby, aggregate_across_components, at_port, bus_carrier, carrier, nice_names, drop_zero, round)
    198 except IndexError:
    199     continue
--> 200 vals = func(n, cn, port_suffix)
    201 if self._aggregate_components_skip_iteration(vals):
    202     continue

File open-tyndp/.pixi/envs/open-tyndp/lib/python3.12/site-packages/pypsa/common.py:705, in pass_empty_series_if_keyerror.<locals>.wrapper(*args, **kwargs)
    702 @functools.wraps(func)
    703 def wrapper(*args: Any, **kwargs: Any) -> pd.Series:
    704     try:
--> 705         return func(*args, **kwargs)
    706     except (KeyError, AttributeError):
    707         return pd.Series([], dtype=float)

File open-tyndp/.pixi/envs/open-tyndp/lib/python3.12/site-packages/pypsa/statistics/expressions.py:1587, in StatisticsAccessor.opex.<locals>.func(n, c, port)
   1585         var = p * p if cost_type == "marginal_cost_quadratic" else p
   1586         opex = var * cost
-> 1587         term = self._aggregate_timeseries(opex, weights, agg=groupby_time)
   1588         result.append(term)
   1590 mapping = {
   1591     "start_up_cost": "start_up",
   1592     "shut_down_cost": "shut_down",
   1593     "stand_by_cost": "status",
   1594 }

File open-tyndp/.pixi/envs/open-tyndp/lib/python3.12/site-packages/pypsa/statistics/abstract.py:122, in AbstractStatisticsAccessor._aggregate_timeseries(cls, obj, weights, agg)
    119         weights = weights / weights.sum()
    120     agg = "sum"
--> 122 return cls._aggregate_with_weights(obj, weights, agg)

File open-tyndp/.pixi/envs/open-tyndp/lib/python3.12/site-packages/pypsa/statistics/expressions.py:375, in StatisticsAccessor._aggregate_with_weights(df, weights, agg)
    373     if isinstance(weights.index, pd.MultiIndex):
    374         return df.multiply(weights, axis=0).groupby(level=0).sum().T
--> 375     return weights @ df
    376 # Todo: here we leave out the weights, is that correct?
    377 return df.agg(agg)

File open-tyndp/.pixi/envs/open-tyndp/lib/python3.12/site-packages/pandas/core/series.py:3266, in Series.__matmul__(self, other)
   3262 def __matmul__(self, other):
   3263     """
   3264     Matrix multiplication using binary `@` operator.
   3265     """
-> 3266     return self.dot(other)

File open-tyndp/.pixi/envs/open-tyndp/lib/python3.12/site-packages/pandas/core/series.py:3237, in Series.dot(self, other)
   3235 common = self.index.union(other.index)
   3236 if len(common) > len(self.index) or len(common) > len(other.index):
-> 3237     raise ValueError("matrices are not aligned")
   3239 left = self.reindex(index=common, copy=False)
   3240 right = other.reindex(index=common, copy=False)

ValueError: matrices are not aligned