Hi, We have an automatic report creation tool that creates Spark SQL jobs based on user instruction (this is a web application). We'd like to give the users an opportunity to visualize the execution plan of their handiwork before them inflicting it to the world.
Currently, I'm just capturing the output of an `explain` statement and displaying it, but it's still too cryptic for users, so we'd love to have something similar to the SQL tab in the Spark UI; from my research, the beautiful SVG displayed there is built from a dot file which is built from aggregated metrics, which are gathered when the job is executing. This appears not to be what I need. Since I'm just getting familiar with Spark's extremely complex codebase, do you guys have any pointers or ideas on how can I provide a more user friendly physical execution plan to my users? Regards, Leonardo Herrera --------------------------------------------------------------------- To unsubscribe e-mail: dev-unsubscr...@spark.apache.org
