At Progressive, our analysts use a software tool called Dataflux for Data Quality analysis. They connect to the hadoop data using the Horton Works ODBC driver. Often times, they complain about slow responses from the Application when they are developing against the Hadoop data. It seems like their queries are executing behind the scenes to get metadata. We have actually seen the query being executed on the hive server even though the developer has only opened their program for development (not execution). Is there a way to stop this from happening? I saw an option to turn on Fast Sql Prepare...but I don't know enough about it and I believe the analysts still need some meta data for their development...does this prevent that from happening? Anyone have any ideas?
... View more