Classification concordances

Look up standard classification codes for the native labels in your database.

For a GLORIA release 060 SUT already parsed as db:

codes = db.get_concordances(
    set="Activity", classification="NACE Rev. 2.1",
    item="Growing wheat",
)
# {'Growing wheat': ['01.11']}

Result · {native_item: [classification_codes]}. Codes remain strings, including leading zeros. The lookup uses packaged metadata and leaves model values unchanged.

Set, classification and item#

Activity or Commodity for SUTs. Sector for IOTs, following the native industry or product identities.

Inspect db.available_concordances. Use an explicit edition such as NACE Rev. 2.1 or CPA 2.2.

One label or a list of labels. Omit it to query the full set. Use db.get_index() to find labels.

Descriptions and provenance#

detail = db.get_concordances(
    set="Activity", classification="NACE Rev. 2.1",
    item="Growing wheat", details=True,
)["Growing wheat"]
detail["matches"]["01.11"]["name"]
detail["matches"]["01.11"]["routes"]

A correspondence is not an allocation rule. Several linked codes do not specify how to split production or emissions. Partial overlaps and composed bridges retain their qualifications in the detailed result.

Supported editions and set defaults

db.table_type
db.sets
db.available_concordances

Which set to query#
Model	`set`	Target normally describes
SUT	`"Activity"`	Activities: NACE or ISIC
SUT	`"Commodity"`	Products: CPA, CPC or the supplied HS correspondence
IOT	`"Sector"`	The native side: activities for industry-by-industry, products for product-by-product

For a SUT, omitting set selects Activity; for an IOT, it selects Sector. The default target is NACE Rev. 2.1 for activity identities and CPA 2.2 for commodity identities. Use explicit editions in reproducible analyses.

Recognised targets; availability depends on the native variant#
System	Explicit editions	Short alias selects
NACE	`"NACE Rev. 2.1"`, `"NACE Rev. 2"`, `"NACE Rev. 1.1"`	`"NACE"` → 2.1
ISIC	`"ISIC Rev. 4"`, `"ISIC Rev. 3.1"`	`"ISIC"` → 4
CPA	`"CPA 2.2"`, `"CPA 2.1"`, `"CPA 2008"`, `"CPA 2002"`	`"CPA"` → 2.2
CPC	`"CPC 2.1"`	`"CPC"` → 2.1
HS	`"HS 2007"`	`"HS"` → 2007

HS 2007 comes from the supplied EXIOBASE product correspondence and its explicit product bridges. A request for HS 2022 raises WrongInput. An available activity concordance does not establish a product concordance.

Region-specific classifications and IOT identities

Without region, a lookup uses the distinct union of native labels across regions. This does not mean every label exists in every region. Restrict the lookup using a region code returned by the model:

region = db.get_index("Region")[0]
labels = db.get_index("Activity", region=region)
regional_links = db.get_concordances(
    set="Activity", classification="NACE Rev. 2.1",
    region=region, item=labels[:2],
)

This matters for BONSAI, whose regional classifications vary. Requesting an item absent from the chosen region raises WrongInput even if another region contains it.

On an IOT, set="Activity" or set="Commodity" is accepted only when compatible with its native Sector identities. SUT-to-IOT conversions retain which side became Sector. BONSAI HIOT also has an explicit reference-product/ process association: an Activity query or a NACE target can follow that bridge. Keys remain the actual HIOT Sector labels; this lookup does not convert the product table into an industry table.

Descriptions, sources and route qualifications

detail = db.get_concordances(
    set="Activity", classification="NACE Rev. 2.1",
    item="Growing wheat", details=True,
)["Growing wheat"]

detail["codes"]                        # ['01.11']
detail["matches"]["01.11"]["name"]      # Target description.
detail["matches"]["01.11"]["routes"]    # Steps and source locations.
detail["sources"]                      # Publisher and source records.
detail["profile"]                      # Selected native variant.
detail["as_of"]                        # Packaged metadata snapshot date.

A route records steps, source locations and qualifications, with traversal direction for composed bridges. Current NACE 2.1 and CPA 2.2 links reproduce the explorer’s reviewed code lists. Other editions use explicit route prefixes or compose official bridges; composed routes are marked partial. Supplied broad anchors stay at their supplied level.

Here a correspondence footprint is the set of classes linked by the source. It can include partial overlaps; it does not assert that the entire native item is exactly partitioned into those classes. Several codes supply no rule for dividing output or emissions among them. Consult route qualifications when using those links to interpret economic or environmental results.

Empty results and unavailable queries

Interpret an empty or detailed result#
Result or `status`	Meaning
`[]` in the simple output	The current native label has no verified link in this lookup. A label absent from the model instead raises `WrongInput`.
`"mapped"`	At least one route exists; source qualifications still apply.
`"unverified"`	Native identity recognised, but no link to this target.
`"unknown_native_identity"`	No bundled identity recognised; possible after renaming, adding or aggregating items.
`"ambiguous_identity"`	The label matches several bundled identities.

Unknown families, unsupported native variants, unavailable target editions and invalid items or regions raise WrongInput. Inspect db.database_profile["concordances"] for family support and variant reasons. Missing links are not filled by fuzzy matching.

To group these codes into sections or divisions, continue with Aggregate by classification.

Continue with Aggregate by classification.