We’re currently on Constellation 24.2 and running a spike to figure out the best way to create a custom search in the portal.
Because we are working with Constellation, we know we’re pretty much limited to using a List View for this.
Our main challenge: We need this List View to be completely empty on load. We don’t want it pulling any records until the user actually types their criteria into the promoted filters and search.
I’d love to get some advice from anyone who has tackled this in Constellation:
Empty on Load: What is the best practice in Constellation to ensure the List View doesn’t run its query until the promoted filters actually have values?
Compound Searches: Since we are passing these promoted filter values down to the data page, what’s the cleanest and most performant way to handle the compound searches (like checking three different phone number columns for one input)?
Any advice, workarounds, or shared pain points would be hugely appreciated.
In Constellation, List Views are the correct mechanism for custom search, and the best-practice way to keep the List View empty on load is not by controlling the UI, but by gating execution in the underlying parameterized List Data Page. By promoting filters to data page parameters and adding defensive logic in the data source (Report Definition or Data Transform), the data page can return zero results and avoid executing any query when all search parameters are blank. This ensures optimal performance, clean design, and upgrade safety. Promoted filters should remain stateless and optional, allowing Constellation to automatically refresh the List View only when criteria are entered. For compound searches (for example, checking multiple phone number columns from a single input), the recommended approach is to handle this declaratively in the Report Definition using conditional OR logic, which is transparent, maintainable, and performant when properly indexed. For very large data volumes, a normalized data model or canonical search column may be considered for additional optimization. Overall, the guiding principle in Constellation is to keep the UI lightweight and declarative, push all conditional logic to the data layer, avoid client-side or imperative workarounds, and rely on parameterized data access patterns that are fully aligned with Pega guardrails and long-term maintainability.