CSV conventions
CSV conventions
BusDK’s preferred default representation for workspace datasets is UTF-8 CSV with a header row and one record per row, paired with explicit Table Schemas. Conventions prioritize compatibility: UTF-8 encoding; comma delimiters; quoting for fields containing commas or newlines; ISO date formats (YYYY-MM-DD) for date fields; and predictable numeric formats for monetary values. The intended result is that the canonical dataset remains readable with general-purpose tooling and straightforward to inspect and diff, consistent with the long-term accessibility goals described in Plain-text CSV for longevity and aligned with National Archives guidance on selecting sustainable formats for electronic records.
CSV is a delivery convention rather than the goal. Other storage backends may be used over time as long as BusDK preserves deterministic, schema-validated tables and predictable export back to simple, tabular text formats.
Audit-trail CSV requirements
To satisfy Finnish audit-trail expectations, tabular datasets MUST be deterministic, linkable, and stable over time. See Finnish bookkeeping and tax-audit compliance.
Each dataset MUST include a stable primary identifier column (*_id) that never changes once written, and cross-dataset references MUST be explicit as foreign key columns (for example: voucher_id, entry_id, attachment_id, bank_txn_id). Row ordering MUST be reproducible using date + sequence columns so that independent tools can produce the same time-ordered view.