Not sure whether better do ask this in an Azure or Spark subject, but I thought I might get responses appropriate to our use cases here.
We have Azure Databricks set up and working, and not had any problems following along the tutorials, but I don't feel they really let me know how to use in practice. I would appreciate any recommendations, but a couple of questions to give an example of the kind of thing I don't know.
When I close a cluster down, I get a warning that all data will be lost. However, the workspace seems to remember that the blob storage was mounted and if I rerun the full notebook I get an error at this step: Should I be separating my mounting ste https://omegle.onl/ ps into a different notebook? Or should I unmount at the end of a notebook? Is this mounted for the entire subscription, or just me as a user?
I have successfully written out last dataframes as a parquet in blob storage. Databricks splits it into many different files. How do I keep track of t https://vshare.onl/ his? Do I need a different folder for each output, or do I rely on databricks to know this? Will this knowledge be retained if I close the cluster? Is there an easy way to view the contents of a folder from within databricks where it displays as a single file; i.e. Output.parquet rather than the 20+ files I see in the actual Blob container.
Any thoughts appreciated.
