The Predicate Push Down feature restructures the way the BriteData report engine queries data to optimize report run times. Reports can take a long time to run due to inefficient querying. This back-end implementation optimizes querying by reducing the amount of data needed for the report, thus requiring less time and memory. Prior to this feature, the report engine would retrieve all data and apply the report filters at the end.
The benefits of Predicate Push Down don't extend to or impact:
- Custom or hardcoded reports. You can tell a report is a custom report if it has a Report Runner of BriteData Py Reports. A hardcoded report won't open to the BriteData editor when Edit is selected. Custom and hardcoded reports may be addressed in the future.
- Dataframe building process. A BriteData report can still be slower if the dataframes aren't present and have to be rebuilt.
- We can't guarantee that a report duration will be reduced by X%. In testing, we have some reports complete with 98% reduction in duration (hours to seconds or minutes) and some reports with less than 10% reduction (one minute to 50 seconds).
No configuration is necessary to take advantage of Predicate Push Down—it's automatically enabled for BriteData Reports.