Diagnosing Empty Simulation Data After Feature Extraction
Summary
This brief BigQuant support exchange addresses a simulation error reported as having no data remaining after missing rows are dropped. The user had already tried extending the feature-extraction lookback by several different periods, without resolving the issue. The proposed diagnostic is to inspect the start date configured in the prediction-set code-list module and check whether it is linked to a live-trading parameter. If that binding causes simulation to request only one day of data, the feature pipeline may not have enough history to produce usable rows.
The exchange therefore points to a configuration mismatch between live parameters and simulation history as a possible cause, rather than assuming that the feature lookback alone is too short. It is a narrow troubleshooting suggestion, not a complete guide: no reproducible setup, follow-up confirmation, or broader diagnosis is given. The explanation should be treated as a hypothesis to check against the project’s date settings and available data.
Key ideas
- An empty dataset after dropping missing rows can result from insufficient history reaching the feature pipeline.
- Increasing the feature lookback may not help if the prediction-set date range is constrained elsewhere.
- A live-trading parameter bound to the prediction-set start date may restrict a simulation to one day.
- The exchange offers a diagnostic possibility but does not confirm that it resolved the reported error.
Tags
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.