Notebooks are nice, but REPL is sometimes more useful. Am I somehow able to run spark-shell that executes on Databricks? Like:
spark-shell --master https://adb-5022.2.azuredatabricks.net/
I looked through available tools related to Databricks (databricks connect, dbx, ...), but it seems there's no such functionality.
CodePudding user response:
Databricks connect is the tool that you need if you want to execute code from you local machine in the Databricks cluster. Same as the spark-shell
, the driver will be on your local machine, and executors are remove. The databricks-connect
package installs the modified distribution of the Apache Spark so you can use spark-shell
, pyspark
, spark-submit
, etc. - just make sure that that directory is in the PATH
.
P.S. but I really don't understand why notebooks doesn't work for you - spark-shell
doesn't have any superior features compared to them.