Reproduce a SAF supply share with one airport table first
Start with an unduplicated numerator and denominator covering the same airport, period and supply boundary.
Before drawing a percentage
Reproducing a sustainable aviation fuel share requires total fuel and SAF within the same boundary. Deciding which batches belong comes before dividing two numbers. Truthound can check the prepared supply table's keys and a caller-defined calculation relationship.
A 2026 report on 2025 supply
EASA published its ReFuelEU Aviation annual technical report on September 17, 2026. Reported 2025 supply was 39.3 million tonnes of aviation fuel, including 1.1 million tonnes of SAF, or about 2.8%. The publication year and the operating year are different.
A deliberately separate example
Keep the first fixture separate from actual airport results. An illustrative 100-tonne total containing 3 tonnes of SAF gives a 3% share. Adding SAF again to create a 103-tonne denominator changes the definition. The fixture tests inclusion; it is not a performance claim about a particular airport.
How one batch arrives twice
Define a business key such as airport, supplier, batch and year. Supplier aggregates and airport receipts can describe the same underlying batch. Combining them as separate contributions may count it twice. Decide which source is for reconciliation and which is for aggregation instead of merely labelling two copies differently.
Supply the calculation meaning to Truthound
Truthound Core Python SDK 3.3.4, operating on a prepared Polars table, provides MultiColumnUniqueValidator for a supplied composite key and ExpressionValidator for caller-defined Polars row expressions. These can check a total/part relationship or an agreed ratio formula. Nulls, denominators of 0, rounding and mass units require explicit policies before the expression is meaningful.
Compare the existing predicate
An existing pandas table with Pandera 0.35.0 user predicates can remain. Compare cases containing SAF above the total, conflicting units and a repeated batch, examining where each is held. Documentation alone does not show that the two paths produce identical values or exceptions. The implemented input contract is the comparison.
Do not add airport percentages
Do not add airport percentages or average them with equal weight by default. A wider supply share needs its numerator and denominator re-aggregated over the intended total population. Preserve which airports are absent and whether both sides cover the same reporting period.
A changed boundary needs another fixture
Revisit the fixture if units or batch boundaries change. A new file may describe supplier reporting rather than actual airport receipt. Keep that source definition visible. A precise-looking percentage can still combine incompatible observation boundaries.
Earn the next airport
Expand when the first table's duplication and inclusion rules are explainable, not because it reports a high pass rate. Add comparable scope first and treat different boundaries as another experiment. That keeps evidence of supply progress separate from the mechanics used to calculate it.
Conditions
Verify mass units and airport, supplier and batch scope for lawfully obtained supply data.
Alternatives
Keep an existing Pandera 0.35.0 pandas calculation check when it uses the same supply boundary.
Limitations
A valid table ratio does not authenticate fuel or certify regulatory compliance.
Evidence
- Event checked
- Subject checked
- Modified
