Thanks JamesY.

Maybe I've misunderstood, but isn't that the goal? To filter out data that will cause the load to SQLMI to fail?

If you're just looking to identify the records that may fail upon write to SQLMI could we just modify the `where` clause as follows:

spark.read.format("csv").option("header", "true").load("/path/to/csvs").where(length("the_column_in_question") > 1024)