slowly changing dimension integrity derived

{
  "input": {
    "body": {
      "csv": "customer_id,valid_from,valid_to,is_current\nC-001,2026-01-01,2026-06-01,false\nC-001,2026-05-01,,true\nC-002,2026-01-01,2026-03-01,false\nC-002,2026-04-01,,true\nC-003,2026-01-01,,true\nC-003,2026-02-01,,true\nC-004,2026-01-01,2026-04-01,false\nC-004,2026-04-01,,true\n",
      "key": [
        "customer_id"
      ],
      "valid_to": "valid_to",
      "valid_from": "valid_from",
      "current_flag": "is_current"
    },
    "type": "http",
    "method": "POST",
    "bodyType": "json"
  },
  "output": {
    "type": "json",
    "example": {
      "keys": [
        {
          "key": "C-001",
          "versions": 2,
          "last_from": "2026-05-01",
          "first_from": "2026-01-01",
          "current_rows": 1,
          "open_intervals": 1
        },
        {
          "key": "C-002",
          "versions": 2,
          "last_from": "2026-04-01",
          "first_from": "2026-01-01",
          "current_rows": 1,
          "open_intervals": 1
        },
        {
          "key": "C-003",
          "versions": 2,
          "last_from": "2026-02-01",
          "first_from": "2026-01-01",
          "current_rows": 2,
          "open_intervals": 2
        }
      ],
      "limits": {
        "maxRows": 5000,
        "maxListed": 200,
        "maxColumns": 64,
        "maxCsvBytes": 262144,
        "maxDatasets": 10,
        "maxFindings": 500,
        "maxCategories": 1000,
        "maxKeyColumns": 4
      },
      "passed": false,
      "coherent": false,
      "complete": true,
      "findings": [
        {
          "key": "C-001",
          "row": 3,
          "code": "INTERVALS_OVERLAP",
          "detail": "C-001: row 2 ends at 2026-06-01 and row 3 starts at 2026-05-01, so both are valid at once."
        },
        {
          "key": "C-002",
          "row": 5,
          "code": "INTERVAL_GAP",
          "detail": "C-002: row 4 ends at 2026-03-01 and the next starts at 2026-04-01, so there is a period with no valid version."
        },
        {
          "key": "C-003",
          "row": 6,
          "code": "MULTIPLE_OPEN_INTERVALS",
          "detail": "C-003 has 2 intervals with no end, so which version is live is ambiguous."
        }
      ],
      "gap_count": 1,
      "key_count": 4,
      "row_count": 8,
      "request_id": "example-request",
      "service_id": "slowly_changing_dimension_integrity",
      "assumptions": [
        "Interval bounds are compared as text, which is correct for ISO-8601 dates and timestamps and is why those are the formats this expects.",
        "An interval is treated as half-open: a row ending exactly where the next begins is contiguous rather than overlapping, which is the convention that makes a point-in-time lookup return one row.",
        "An end equal to the declared open-end value, or empty, means the version is still live."
      ],
      "limitations": [
        "Temporal coherence within the dataset supplied. Whether the values in each version are correct is not assessed.",
        "coherent is true only when the dataset was fully read and every key had exactly one live version, no overlap and no gap."
      ],
      "observed_at": "2026-10-02T00:00:00.000Z",
      "not_examined": [],
      "finding_count": 5,
      "overlap_count": 2,
      "open_end_value": "",
      "service_version": "1.0.0",
      "findings_by_code": {
        "INTERVAL_GAP": 1,
        "INTERVALS_OVERLAP": 1,
        "MULTIPLE_CURRENT_ROWS": 1,
        "INTERVAL_OVERLAPS_OPEN": 1,
        "MULTIPLE_OPEN_INTERVALS": 1
      }
    }
  }
}