I have a dataset (D2) based on another dataset (D1) that has raw reflection turned on. When selecting D2 from the UI it automatically executes in preview mode (haven’t figure out how to turn that off). That query never gets accelerated because it says the D1 reflection doesn’t cover the query. However, when I select RUN to execute the query it does use the D1 reflection. The net result is the preview takes much longer to execute then just running the query.
Are there any use cases where this is the expected behavior?
Can I turn off preview? It’s really annoying that it takes longer to preview then to run my query. I’d just as soon just open the page rather then automatically run any query in addition.
This is just me learning Dremio so tiny dataset. Still the preview typically takes 24 seconds to run while just executing the accelerated query is typically less then 2 seconds. (single node dremio on my local mac book pro). The source is parquet files in S3.
Hey @swarren another option is decreasing the sample size we use for previews. You can configure this option by going to Admin > Advanced Settings > Dremio Support and entering this key: planner.leaf_limit_size. Defaults to 10000 records per thread.
Roger that, will do. I will still suggest the ability to turn off automatically running any query when navigating to a dataset. It is annoying especially if your main reason to navigate to any dataset is to change the query.