Power Automate Best Practices (Working With Data)

Power Automate flows often struggle with large datasets. Unnecessary Apply to each loops process every item individually. This increases run times exponentially, raises licensing costs, and risks timeouts in production.

Jon Manderville covers three essential techniques to handle data selection and filtering efficiently. Attendees learn to implement OData filtering, Select actions, and Filter Array actions. These approaches minimize loops and deliver speedups of 20 times or more for flows working with bigger data volumes.

This summary is based on the session title and description.

What You Will Learn

• How to identify unnecessary Apply to each loops in existing flows and refactor them for better performance
• Setting up OData filtering queries directly in connectors like SharePoint, Dataverse, or SQL to retrieve only required records
• Using the Select action to map, rename, and transform array data without iterating over each item
• Applying Filter Array actions to narrow down arrays based on conditions after initial data retrieval
• Combining OData, Select, and Filter Array to process large datasets end-to-end without loops
• Measuring flow performance using run history analytics to validate 20X or greater reductions in execution time
• Best practices for error handling and fallback strategies when filtering large data sources

Key Takeaways

1. Filter at the source with OData queries. Build conditions into your connector steps to avoid pulling entire lists into Power Automate, which eliminates the need for initial Apply to each loops.
2. Leverage Select for data transformation. Map fields from complex objects to simple arrays in one action, preparing data for downstream steps without processing items one by one.
3. Use Filter Array for client-side refinement. After minimal data retrieval, apply conditions to arrays to keep only relevant items, often cutting processing time by orders of magnitude.
4. Profile loops in run history first. Check loop iteration counts and durations to target optimizations, then retest to confirm gains like 20X faster runs on production datasets.

About the Speaker

Jon Manderville presents Power Automate sessions for Collab365, the Microsoft 365 community platform. He specializes in practical optimizations for flows handling real-world data volumes. His content targets IT professionals seeking scalable automation solutions.

Who Should Watch This

This session suits Microsoft 365 admins and developers whose flows process lists, tables, or API responses with hundreds or thousands of items. You gain immediate value if current flows use Apply to each on large datasets, leading to slow performance or high costs.

Power users transitioning to production automations benefit from learning these filtering techniques early. Teams managing flow quotas across organizations find the performance tips essential for sustainability.

Experienced flow builders comfortable with expressions may still pick up nuances on OData syntax and array handling. Skip this if your automations stay under 100 items per run, as basic loops suffice there.