Identity CSV Import Validator
Map headers, validate declared types and review duplicate identities, references and manager cycles.
Local schema review only. This does not check provider import rules, live identities or directory state. Review and issue artifacts contain original identity data; inspect before sharing. Clean export requires zero blocking errors and retains all rows.
Maximum 8,000,000 file bytes, 2,000,000 text characters, 10,000 data records, 100 columns and 16,384 UTF-8 bytes per cell. Files are streamed and discarded; an optional UTF-8 BOM is removed during file decoding.
Version 1 schemas declare exact source → target names, type, required/unique rules and optional normalization. Unmapped columns remain unchanged. Types: string, integer, decimal, boolean, date, datetime, multivalue. Integer/decimal JSON values stay exact strings. Dates require YYYY-MM-DD; datetimes require T, seconds and Z or an explicit offset, with up to 9 fractional digits. No guessing, exponent notation or leap seconds.
Normalization: none, trim, lowercase, trim-lowercase. Boolean defaults: true/1 and false/0; custom trueValues/falseValues must be distinct. Multivalue delimiter defaults to semicolon; empty members block export and duplicate members are retained with a warning. reference.column selects a declared scalar target name; manager: true enables cycle checks for one string/integer manager field.