Azure Data Lake Query Speed


  • We're just starting to play around with Azure Data Lake and were a little shocked at the slow speed of running a simple USQL query (ie select * ...where  created > date1 && created < date2 && temp > 85.5 ).  I realize it's designed to crank out massive results, but still wasn't prepared for the long wait.  When we run a query in the data explorer it appears to create a "plan" and then executes it.  Is there any way to cache the "plan" so that the query can be faster when run subsequently?

    We are running in an IOT environment and will need to extract small sections of sensor data periodically for historical replay.  Looks as though we may need to add a DocumentDB in addition to our Data Lake for report retrieval if we can't find a way to get the data quicker.  Trying to avoid that expense!  Any suggestions?

    Tuesday, March 28, 2017 2:12 AM