Exact-key CSV reconciliation with duplicate exceptions

Exact-key export reconciliation — self-initiated Python demonstration with fictional records, not client work.

The problem: duplicate IDs can silently create misleading matches when two exports are joined. This demonstration compares exact textual project IDs and returns matched, changed, left-only and right-only rows plus explicit duplicate-key exceptions.

Verified sample result: one matched record, one changed record, one left-only record, one right-only record, and six duplicate exceptions across two ambiguous IDs. If a key repeats on either side, every occurrence on both sides is excluded from matching. Two automated checks cover duplicate exclusion and the output categories with exact textual-value preservation.

Leading-zero IDs are preserved. Values 12 and 12.0 deliberately differ as strings. Inputs use the demonstrated fixed schema and nonempty keys. Python standard library only; local CSV processing. No fuzzy matching, inferred corrections, financial certification or general production-platform claim. Built with AI assistance.

See the original demonstration summary: https://contra.com/community/wejrYV6q

For a scoped data task, message with column names, required output and fictional examples. No confidential records. Scope, acceptance checks and delivery terms are confirmed before any service order.