Best practices to reduce memory usage of queries

When running queries you might hit Out of Memory errors, preventing your queries from completing and showing visuals in your reports.

Here are some best practices to ensure your setup is more optimized

  • Run compute stats on all your raw tables. This is critical for planning execution of queries and optimizing resource usage

  • Partition large tables and use them to filter visuals as much as possible. This will allow you to leverage partition pruning

  • Materialize complex joins. Denormalizing into a single table will help avoid computing joins repeatedly on the fly. Instead push to ETL so the join is computed only once.

  • Leverage Analytical Views. Use the recommendation engine as explained here to build acceleration artifacts for predictable workloads.
    How can I enable "Smart Acceleration" in Arcadia to get Analytical View recommendations for my Dashboards?

1 Like