DATASET QUALITY CHECKER

Find the problem rows before importing a batch

One missing field can hide in a batch. Check FakeKit fictional records together, see which rows need attention and export an issue report without copying the input values into it.

FICTIONAL QA ONLY Local processing · No accounts · No uploads

Input formatReference date

Expiry and future-date checks use this date, not your device clock.

Fictional test records

[
  {
    "recordId": "TEST-0001",
    "classification": "FICTIONAL TEST DATA — NOT VALID",
    "validForIdentification": false,
    "documentKind": "document",
    "fullName": "Sample Person 0001",
    "issuer": "Fictional Test Authority",
    "documentNumber": "NOT-VALID-0001",
    "dateOfBirth": "1990-01-01",
    "issueDate": "2020-01-01",
    "expiryDate": "2040-01-01",
    "scenario": "standard"
  },
  {
    "recordId": "TEST-0002",
    "classification": "FICTIONAL TEST DATA — NOT VALID",
    "validForIdentification": false,
    "documentKind": "document",
    "fullName": "",
    "issuer": "Fictional Test Authority",
    "documentNumber": "NOT-VALID-0002",
    "dateOfBirth": "1990-01-01",
    "issueDate": "2020-01-01",
    "expiryDate": "2040-01-01",
    "scenario": "standard"
  },
  {
    "recordId": "TEST-0001",
    "classification": "FICTIONAL TEST DATA — NOT VALID",
    "validForIdentification": false,
    "documentKind": "document",
    "fullName": "Sample Person 0003",
    "issuer": "Fictional Test Authority",
    "documentNumber": "NOT-VALID-0003",
    "dateOfBirth": "1990-01-01",
    "issueDate": "2020-01-01",
    "expiryDate": "2040-01-01",
    "scenario": "standard"
  }
]

1–200 records, up to 128 KiB. FakeKit common-field schema only. Do not paste personal data.

Dataset reportNOT AUTHENTICATION

The example includes a missing name and a duplicate record ID. Run the check to inspect them.

A report will appear here.

It includes affected rows, issue fields and duplicate test IDs. Input values are omitted from report exports.

A narrow, useful check

What this report checks
Check	Reported behavior
Common fields	Required strings, supported kind and permanent fictional markers
Date fields	Calendar validity, birth/issue order, issue/expiry order and reference-date limits
Identifiers	TEST- and NOT-VALID- formats, plus duplicate record IDs across the batch
Result counts	Rows passing these rules, affected rows and per-field affected-row counts

The checker uses the same common-field rules as the single-record validator. Optional passportTest and licenseTest objects are not validated. Unknown extra fields are not a schema error, and a rule pass does not authenticate a document or verify a person.

JSON and CSV have different boundaries

JSON input must be an array of 1–200 records. CSV must contain a header followed by records, with at most 30 distinct simple column names and a consistent column count. Quoted commas, double quotes and line breaks are supported. The exact CSV value false in validForIdentification is converted to a boolean; other cells remain strings.

Each input is limited to 128 KiB of UTF-8 text. A malformed CSV stops the batch so row boundaries are not guessed. A malformed JSON row is reported as a row issue. An individual JSON record exceeding the single-record validator's size limit is reported without being validated.

Read the report before acting

Affected-row counts are not error totals: a row can have several issues, and one field is counted once per affected row. All occurrences of a duplicate record ID are flagged. Row numbers refer to records; CSV row 1 in the report is the first row after the header.

The page shows the first 20 affected rows. JSON and CSV report exports include every row result, omit original field values and retain their non-authentication label. Inputs and reports remain in page memory until you leave or reload; downloaded files remain on your device.