A cluster is a named group of members of one set. Store a selection once, then reuse it in workflows that accept clusters, such as shocks or sector additions.

Create a cluster

import mario

db = mario.load_test("IOT")
sectors = db.get_index("Sector")
db.set_clusters({"Sector": {"Selected sectors": sectors[:2]}})
selected = db.available_clusters["Sector"]["Selected sectors"]

Result · A reusable named selection containing the first two native sector labels. This call does not change the table’s dimensions or values.

Set names accept the usual aliases; members must exist in the chosen set. See db.set_clusters() for validation and replacement options.

Default and saved groups

Generated regional groups

db.default_clusters

On recognized region labels, MARIO derives groups such as continents and EU membership. Default groups are generated only for Region; their members depend on the database’s coverage.

All available groups

db.available_clusters

Combines generated defaults with the groups you have saved. You can define your own clusters for other sets, such as Sector.

For example, on a parsed EXIOBASE database:

db.default_clusters["Region"].keys()
eu_members = db.available_clusters["Region"]["EU"]

The artificial region labels in the packaged test table do not identify countries and cannot provide this geographic example.

Clusters and aggregation

A cluster records a reusable selection. Aggregation changes the table and needs one target group for each original member being mapped.

The aggregation guide explains how to use regional presets (continent, UNregion, EU, OECD, G7, G20) or a custom mapping. Region presets share the coverage logic used by default clusters; selecting a cluster alone does not aggregate anything.

Continue with shocks or db.available_clusters.