DLI Spark does not support job scheduling. You can use other services, such as DataArts Studio, or use APIs or SDKs to customize job schedule.
The Spark SQL syntax does not support primary key definition.
Yes.
DLI built-in dependencies are provided by the platform by default. In case of conflicts, you do not need to upload them when packing JAR files of Spark or Flink Jar jobs.
No, the packages cannot be downloaded.