Tuesday, May 14, 2024
1 change · saas-17.2
Enhancements to existing features
Spreadsheet documents now load only the revision information needed for the document overview instead of pulling in large revision histories. This reduces unnecessary data loading and helps large spreadsheets appear faster in document lists.
Original PR description
Since a055dfeb84ab, the field `current_revision_uuid` is in the documents kanban view. Which means it can be computed for many documents. The compute method loads all the revisions of spreadsheet documents. Revisions can hold a lot of data (field `commands`), several hundreds of Mb for large spreadsheets. With this commit, we pre-fetch the only field that we need to prevent the ORM to prefetch all fields that we don't need (and specifically avoid loading the field `commands`) Task: 3918564