AX-R03 — Declared largest single impurity exceeds declared total
A certificate reports a single impurity larger than the sum of all impurities.
The predicate
Published literally. A rule you cannot read is an accusation you cannot check.
coa.largestSingle != null && coa.totalImpurities != null && coa.largestSingle > coa.totalImpurities + 0.02
Why this rule exists
This is arithmetically impossible and there is no innocent interpretation involving the measurement. The largest single impurity is a member of the set whose sum is the total. It cannot exceed it.
Every instance the index holds resolves to one of two clerical failures. Either the two fields were transposed on the certificate template, or the "total impurities" field was populated with the largest single value and the "largest single" field with something else entirely. Both are template errors rather than data errors, and in most cases the underlying chromatogram, when requested, was fine.
The index publishes the flag anyway because a reader reconciling a certificate against a record needs to know which of the two fields to trust. In every instance the index has followed up, the underlying integration was supplied on request and the certificate reissued with the fields the right way round. Asking for the related-substances table rather than the summary field avoids the ambiguity altogether, and every supplier scorecarded here will send one.
Incidence
How often the rule fires, where, and when. A rule that never fires is not calibrated; a rule that fires on half the index is measuring the population rather than screening it.
| Supplier | Flagged | Records in index | Rate | |
|---|---|---|---|---|
| SGN | 22 | 221 | 10.0% | |
| JEEP | 15 | 230 | 6.5% | |
| GGPeps | 14 | 371 | 3.8% | |
| WXT | 10 | 301 | 3.3% | |
| TFS | 10 | 188 | 5.3% | |
| BCH | 9 | 209 | 4.3% | |
| SWB | 5 | 204 | 2.5% | |
| PPG | 3 | 213 | 1.4% | |
| SIG | 2 | 160 | 1.3% | |
| CPC | 2 | 227 | 0.9% | |
| TFC | 1 | 417 | 0.2% |
Flagged records per quarter
2024-Q1: 0 · 2024-Q2: 0 · 2024-Q3: 16 · 2024-Q4: 12 · 2025-Q1: 10 · 2025-Q2: 15 · 2025-Q3: 10 · 2025-Q4: 0 · 2026-Q1: 13 · 2026-Q2: 17 · 2026-Q3: 0