For the SQL kernel, we currently attempt to get all of the rows from a resultset in a synchronous manner. This has perf implications when resultsets get larger and larger, so we need to converge on a nice user experience here. For example, we could try to only get more rows from the resultset when a user starts to scroll.
In addition, we need to be mindful of the user experience for saving a notebook; we need to be explicit in how many rows we serialize, and surface that to the user.
The text was updated successfully, but these errors were encountered: