Is Data Fix meant for correcting all my meter data or just small issues?
Data Fix is designed for targeted, minor corrections — not for rewriting large datasets. Its purpose is to patch small data quality issues that occur naturally during meter data collection, such as a handful of missing timestamps, a brief negative reading glitch, or a short flat plateau caused by a communication dropout.
What Data Fix is designed for
- Minor negative values — A cumulative meter briefly dips below its previous reading due to a rounding error or sensor glitch (e.g., 5000 → 4998 → 5003). Data Fix corrects the single invalid point.
- Short flat plateaus — A meter reports the same value for 10–48 hours because communication was interrupted, then a spike arrives with accumulated consumption. Data Fix redistributes the spike across the flat period.
- Missing timestamps — A few readings are absent in an otherwise continuous stream. Data Fix interpolates estimated values to fill gaps up to 7 days by default.
- Duplicate readings — Two readings arrive at the same timestamp due to a re-transmission. Data Fix merges them.
- Brief out-of-range spikes — A gauge meter reports an impossible value for one or two readings. Data Fix caps it to the expected range.
What Data Fix is NOT for
- Replacing weeks or months of missing data — if your meter was offline for an extended period, the root cause needs to be fixed (reconnect the meter, repair hardware, reconfigure the API connection).
- Overriding correct readings that you disagree with — Data Fix targets anomalies, not valid consumption you consider too high or too low.
- Bulk data migration or reformatting — use Data Import for loading data from a new source.
- Correcting systematic errors in your data pipeline — if every reading is wrong, the import configuration or meter hardware needs attention, not a per-reading patch.
Built-in scope limits
- Maximum 90 days per operation (server-enforced)
- Maximum 100,000 readings analyzed per run
- Gaps longer than 7 days are skipped by default
- A warning fires when more than 10% of readings need fixes — this signals a deeper problem
Where to access: Analytics → Data Fix.