This isn’t a very new issues but has got a lot worse since knime 4.4. Namely when I first open a node outputs data table and that datatable contains a chemical structure type the first loading is very, very slow and the slower the more rows there are.
Thanks for reporting this, I don’t think we were aware of the issue.
Is it possible for you to share any example workflow that runs into the issue? You can also send it (with a reference to this post) to email@example.com if that is preferable.
That would make it a lot easier to investigate for development.
Could you also check your exact version number? Either during program start or under Help → About KNIME Analytics Platform.
In 4.5 it’s only the first load after starting KNIME that is slow. From then on it’s now fast regardless if I then select a different table. I use Marvin plugin for rendering. I assume that gets loaded once on first table load?
I assume so. On the columnar backend side we added a way to make so called “BlobCells” load faster and especially avoid redundant loading of these. That’s why it’s faster with 4.5. The initial overhead when loading the table is very likely related to Marvin itself or other “one-time” loading operations (e.g. backend initialization or so).