How to determine if Spark is rewriting data

  • Article

First open the SQL DAG for your write stage. Scroll up to the top of the job's page and click on the Associated SQL Query:

Stage to SQL

You should now see the DAG. If not, scroll around a bit and you should see it:

SQL DAG

If you're doing a Delete or Update operation, look at the amount of data being written by the writer versus what you expect. If you're seeing a lot more data being written than you expect, you're probably rewriting data:

Write Stats

If you're doing a merge, the merge node has explicit statistics about how much data it's rewriting.