v1: solve-time coordinate check rejects models with independent label sets on a shared dimension (breaks PyPSA) #986

Version Checks (indicate both or one)

• I have confirmed this bug exists on the latest release of Linopy.
• I have confirmed this bug exists on the current master branch of Linopy.

Issue Description

Since #955, Model.solve() runs _check_coord_consistency() under semantics="v1". It raises when two containers share a dimension name but their labels do not nest.

PyPSA uses one dimension name for all components (Generator-p, Store-e, Bus-nodal_balance, ...), each with its own disjoint labels. So every PyPSA model with more than one component type fails at solve time under v1. Running PyPSA master tests (test_lopf_storage, test_lopf_multiinvest, test_optimization_expressions, test_components_store) with LINOPY_SEMANTICS=v1 against linopy master gives 42 of 148 failed, all with this error.

These variables are never combined element-wise, so no ambiguous alignment happens. The model is valid.

Reproducible Example

import linopy
import pandas as pd

linopy.options["semantics"] = "v1"

m = linopy.Model()
gen = m.add_variables(lower=0, coords=[pd.Index(["gen1", "gen2"], name="name")], name="Generator-p")
store = m.add_variables(lower=0, coords=[pd.Index(["store1"], name="name")], name="Store-e")
m.add_constraints(gen.sum() + store.sum() >= 1, name="balance")
m.add_objective(gen.sum() + store.sum())
m.solve()

ValueError: Coordinates for dimension 'name' are incompatible across the model: 'Store-e' carries labels not contained in the largest carrier 'Generator-p', so they are neither equal nor subsets of a shared index. Use Model.assign_coords to relabel the model, or align the containers with `.sel`.

Expected Behavior

The model solves, as under legacy semantics. Containers that share a dimension name without being combined should not need identical or nested labels. Alternatively, v1 should document how downstream packages like PyPSA should model independent label sets on a shared dimension name, before the check becomes a hard error.

Installed Versions

• linopy 0.9.1.post1.dev42+g7f06ffaf1 (master, 7f06ffa)
• xarray 2026.7.0
• pandas 3.0.3
• python 3.13.11