The problem with Tableau’s incremental refresh feature is that it still requests all the data from the database, before discarding everything old and keeping everything new. If you get billed for the amount of data you request from the database, or have a complex SQL statement with a lot of calculations that takes a long time to run, or a relatively small amount of memory or disk space on the database server, this can be a big problem.
Why don’t the main R and Python keyring packages, r-lib/keyring and jaraco/keyring, seem to be able to read the same passwords from Windows Credential Manager on Windows 10? In short: because the R package gives the credentials a name like :service-name:user-name, while the Python package simply uses service-name (or user-name@service-name if there are multiple credentials for the same user). If you want to store credentials for user mjw for the service local-database, the best way to do that is to open Windows Credential Manager (Start > Credential Manager), click Add a generic credential, and enter:
All plots generated within an R Markdown document chunk take the width and height defined in that chunk’s options, meaning that all plots within a chunk will be the same size. There is a hacky way to get around this restriction, though.