Colombia Compra Eficiente: secop-providers

Cobertura publicada

Registros

179,202

Última actualización

27 set. 2026

Frecuencia

daily

Frescura

CURRENT

Identificador

nit

Búsqueda por nombre

Disponible

Conteo y frescura consultados desde api.latinfo.dev/sources durante el build del 2026-09-28.

Fuente primaria

• Institution: Agencia Nacional de Contratación Pública — Colombia Compra Eficiente (CCE)
• Data: SECOP II registered-provider profiles, including organization-declared contact channels and procurement classification.
• Format: UTF-8 CSV, comma-delimited, 25 columns.
• Registered: https://www.datos.gov.co/Estad-sticas-Nacionales/SECOP-II-Proveedores-Registrados/qmzu-gj57
• License: Creative Commons Attribution-ShareAlike 4.0 International (CC BY-SA 4.0), as declared in the datos.gov.co metadata.

Cómo se procesa

1. latinfo pipe runs the custom streaming importer on Linux Mint.
2. Gate 1 uses a bounded Socrata API sample; full gates download the official CSV export.
3. The importer retains only values from an explicit juridical-company-type allowlist. Natural persons, OTRO, undefined types, consortia, temporary unions and other ambiguous subjects are excluded.
4. SECOP's inconsistent NIT representations are canonicalized to the zero-padded 10-digit NIT body used by co-rues. A present verifier digit must be valid, and padded placeholders with fewer than six significant digits are rejected.
5. Provider-declared organization channels are projected. Legal-representative name, document, phone and email columns are never written to the canonical TSV.
6. Placeholder values such as No Provisto and No Definido become empty strings.
7. The TSV is sorted by NIT and provider code, then built as a multi-record binary source with V3 name search and an MPHF index.
8. Content-addressed shards and the manifest are uploaded atomically to R2.

The authenticated lookup endpoint is:

GET /co/cce/secop-providers/nit/{nit}

A NIT can return multiple records because SECOP II may contain multiple provider registrations for the same tax entity.

After the semantic profile and vocabulary gates pass, the universal endpoint joins these rows to the RUES identity without replacing the registry legal name:

GET /v1/companies/CO/{nit}

The universal response labels these channels as company-declared to an authority and not independently verified. Ten-digit Colombian phones are normalized to E.164 with +57; shorter numbers remain explicitly local rather than being made artificially dialable.

Campos publicados

Limitaciones conocidas

• Contact channels are declared by the provider and are not independently verified by Latinfo. They may be stale, shared, malformed or hosted on free email domains.
• One NIT can have multiple SECOP II provider codes; the endpoint intentionally returns an array.
• OTRO is the largest non-natural label upstream but is excluded because it does not prove juridical status. It can be added later only through a strong RUES identity match.
• Consortia, temporary unions, autonomous estates and similar non-person structures are excluded by the juridical-only policy.
• Some upstream NITs are malformed or carry an invalid verifier digit and are skipped with aggregate counters.
• Empty upstream placeholders are returned as empty strings.