Normalized Key Collision Reviewer

How to use Normalized Key Collision Reviewer

Before normalizing identifiers for a join or migration, check whether different original values would become the same key. Exact namespaces stay separate, repeated equal keys are distinguished from true normalization collisions, and every contributing data-record number is retained. The tool produces evidence without rewriting or deleting records.

1. Paste a rectangular CSV and name the exact identifier column.
2. Choose up to three namespace columns and explicit normalization/case/edge policies.
3. Review collision groups and original record numbers before changing the source dataset elsewhere.

Good to know

• At most 200,000 well-formed CSV characters, 2,000 data records and 64 unique nonblank headers. Each record must have the same width. Namespace is a JSON array of at most three distinct existing columns, excluding the key itself.
• NFC, NFKC or no normalization runs after optional U+0020 edge-space trimming; optional ASCII A–Z lowercasing follows normalization. No full Unicode case folding, locale rules, internal-space collapse or visual-confusable inference occurs.
• Only groups with two or more distinct original key strings are collisions. Repeated identical strings alone are not. Namespace values remain exact, and the blank policy applies to the resulting normalized key.
• The JSON report preserves original values and 1-based data-record numbers rather than physical CSV line numbers. No rows are edited, merged or removed, and no identifier-security, identity or uniqueness guarantee beyond this closed input is implied.
• Processing is local. Output is limited to 1,000,000 text characters unless a stricter file limit is stated.