Common Conventions

Audit Columns (almost always present)

Most tables inherit standard base fields such as:

• updated_at: last known modification of the record at the source.
• **deleted_at:** logical deletion; a filled value indicates that the record was removed in the application (depending on the extraction, rows may remain for historical purposes).

Use these fields to determine whether a relationship is still active.

Multi-tenant and Scope

Many tables include identifiers such as:

• **tenancy_id:** logical isolation unit (client/business unit within the platform).
• **business_group_id:** grouping of companies under the same economic organization, when applicable.

Corporate reports usually filter by one or the other depending on the data contract.

Keys and Relationships Between Tables

• Names ending with *_id generally refer to other exported tables (e.g., company_id, waste_id). Integrity follows the application rules; there are no physical foreign keys in the analytical layer.
• Some table names are historical or simplified: the catalog indicates when the granularity does not match the name (e.g., the waste_disposals table built from disposal movement records).

Types and Dates

• Date/time fields are typically exposed as dateTime in the Parquet pipeline.
• Numeric values may be null when the source did not provide data; interpret null as “not provided,” not necessarily zero.

Text and Language

• Descriptive fields may be in Portuguese, depending on how they were registered in the platform.
• Supporting tables (e.g., translation decoding tables) exist to harmonize labels when necessary.