Author: Neha Hasija is a Senior Consult and Board Community Captain.
In my project work, I have seen procedure run time become a practical concern when calculations repeat across a large Layout or an export evaluates more combinations than needed. Working through these cases taught me to look for the workload behind the delay before changing the procedure: repeated calculations, the export method, or the number of combinations. The examples below share approaches I have used and the situations where they helped.
Repeated Last Value calculations
When a Layout combines several Last Value functions with algorithms, those calculations can repeat work across many intersections. For date-range logic, I have found it useful to stage the Last Value results in temporary cubes and let the algorithm use those stored results to flag each day between the start and end dates. The diagrams show the difference between calculating directly and staging the results first.
Dataflow pattern: Last Value results feed the date-flag algorithm.
Temporary-cube pattern: Last Value results are stored before the date-flag calculation.
Dynamic arrival-date calculations
When an arrival date depends on a start date and a dynamic lead time, I have used an Export Dataset and Text File Data Reader flow to calculate the date outside a large Layout. In the Data Reader ETL, the YYYYMMDD start-date value is converted to a date and then an integer; the lead time is added, and the result is formatted as YYYYMMDD before it is loaded into a temporary cube.
Procedure pattern: export the result, then reload it with a Text File Data Reader.
ETL example: add the dynamic lead time to the start date, then format the arrival date as YYYYMMDD.
Choose the export method based on what the next step needs
When the next step does not depend on ranking, sorting, or DataView filters, I consider Export Dataset for a straightforward export. In my experience, it can run faster than Export DataView in these cases. When the process needs those DataView features, Export DataView remains the better fit.
Keep Export DataView combinations manageable
When an Export DataView evaluates too many combinations, I first look for ways to reduce the layout and the data it needs to process:
- Remove columns that are not required.
- Keep only the entities needed on rows.
- Apply relevant filters and selections before exporting.
- If the combination count remains high, split the export into smaller exports by a parent entity, such as Stock Group for Site.
In a Cloud environment, the Cloud team can advise whether a modest increase to the configured maximum row limit is appropriate. Increasing the limit too far can affect performance.
These patterns are starting points, not universal fixes. Performance varies by model, data volume, and selections, so compare run time under the same conditions before and after a change. What has helped you reduce procedure run time in your Board projects? Share your experience with the Community.
Related Board references