CoreModel.get_index(index, level='main', *, region=None, flat=False)

Return one index level or the full index mapping for the database.

Parameters:

• index – Set name such as Region or Sector. Pass "all" to retrieve every available set for the table type.

• level – Index variant to return, typically "main" or "aggregated".

• region – Optional region whose members should be returned as a list. Available only for level="main".

• flat – When True, explicitly return the union of distinct labels as a list even for a regional classification. Available only for level="main" and mutually exclusive with region.

Returns:

For level="main", a list when every region has the same members, otherwise {region: list_of_members}. Regions without members have an empty list. Global sets remain lists. With index="all", a dictionary of these results keyed by set name. Other index levels retain their existing representation.

Return type:

list | dict

The usual return value is an ordered list. For the special case of regional classifications, the result is a dictionary of lists:

db.get_index("Activity")
# Common classification: ["A1", "A2", ...]
# Regional classification: {"IT": ["A1"], "DE": ["A2", "A3"], ...}

Each set is checked independently. Equal counts alone do not imply a common classification. Differences in axis order alone do not require a dictionary; the common list retains the existing logical index order. Regions without members appear with an empty list.

Select one region, or explicitly request the union of distinct labels:

db.get_index("Activity", region="IT")  # Always a list.
db.get_index("Activity", flat=True)    # Union, always a list.
db.get_index("all")                    # Each set uses its own representation.

The union does not imply that all its members exist in every region. Use the regional view when validating region-item pairs.

Membership includes existing nodes with zero flows. Non-main levels, including aggregation DataFrames, retain their existing representation.