When I try to run a simple limit number of rows in Spark using a pushdown query,  I receive an error back from spark with a where clause of 1= 0.  Meaning it won't return the results and the original sql query was appended by spark.

This seems very odd that spark won't allow a "SELECT TOP 100 * From Table" or "SELECT * FROM Table Limit 100"  both of these queries return nothing.  It appears spark wants me to bring in the 10 M records then limit them in the data frame.  This is crazy and very taxing on the DB.