Avoid a fourfold error when aggregating quarter-hour electricity trades
Quarter-hour power and energy need interval-aware checks; adding MW values is not an energy total.
A different interval changes the sum
Four quarter-hour trade rows do not turn their power values into an hour's energy simply by being added. Power, energy and interval duration must be defined together. Truthound is best placed at the boundary of that declared time-and-conversion contract, rather than treated as a tool that infers and repairs units.
A monitoring report, not a new launch
ACER's November 4, 2026 electricity-market integration report describes almost 90% of day-ahead electricity trading as using quarter-hour products. Those products were introduced on October 1, 2025. The new event is the monitoring report on their widespread use, not a November launch of the trading interval.
A minimal conversion example
Use an illustrative input, not a claimed market-data result, to test the meaning. A constant 15 MW across four 15-minute intervals represents 3.75 MWh per interval and 15 MWh in total. Adding the four power values gives 60, which becomes a fourfold error if labelled MWh. This arithmetic example explains a failure mode; it does not report a failed operating file.
Not every local day is identical
Daily aggregation adds a timezone question. A local daylight-saving transition day cannot always be treated as a fixed 24-hour period without dropping or repeating intervals. Preserve interval start and end, timezone and the market area's calendar. Derive the expected interval count from the region and date rather than hard-coding 96 for every local day.
Two policies to supply to Truthound
Truthound Core Python SDK 3.3.4, operating on a prepared Polars table, provides TimezoneValidator for datetime timezone metadata and ExpressionValidator for a caller-defined Polars row expression. Checking energy against power and duration still requires a definition of average power, interval boundaries, null handling and rounding tolerance. A timezone check alone cannot establish that the unit conversion is correct.
Compare the meaning of the predicate
Pandera 0.35.0 is an alternative where an existing pandas workflow already maintains the conversion predicate. Compare both paths on an ordinary 15-minute interval, a different interval length, a missing end time and a duplicate interval. Align what the predicates mean first. Their documentation does not establish that either library produces more accurate market aggregates.
A negative value may be meaningful
A table that represents flow direction with a sign cannot reject every negative value as invalid. Establish the import/export convention and unit first. Correct time conversion and meaningful trade direction are separate questions; forcing both into one non-negative range rule would remove legitimate information from the aggregate.
Separate causes instead of auto-correcting
An unexpected sum should not trigger an automatic divide-by-four correction. Converting an input that was already MWh is a different failure from adding MW directly. Preserve original fields separately from derived energy so the team can identify where unit, duration or conversion stage changed.
When to accept the daily total
Accept the daily total only when the intervals form the intended coverage, do not overlap, and have an explainable unit-and-conversion boundary. If any condition is unresolved, hold the aggregate before interpreting it as a price or demand change. Finer market time resolution makes these physical meanings more important, not less.
Conditions
Define average power, energy units, interval boundaries, regional calendar and timezone policy first.
Alternatives
An existing Pandera 0.35.0 pandas predicate can retain the same interval-and-unit contract.
Limitations
A valid conversion expression does not establish the overall completeness or economic correctness of market data.
Evidence
- Event checked
- Subject checked
- Modified
