3 Replies Latest reply on Oct 11, 2018 9:01 AM by Jeff Strauss

    Query performance

    arvin.ramchurn

      When adding this extra filter we drop from very fast <2 seconds to over 7 seconds.

      Can anyone see what we should do it make the query faster or have that data already availble to filter faster.

       

      Thanks in advance

        • 1. Re: Query performance
          Jeff Strauss

          Hi Arvin.  From looking at your perf recording, it appears you're going through sqlproxy (aka dataserver).  Are you connecting up to an extract or a live connection?  If live, do you have a proper index on the field you're trying to filter?  What is the cardinality of the field you're trying to add with the filter?

          • 2. Re: Query performance
            arvin.ramchurn

            Data loaded into Hyper via Tab Prep  - so extract no db. its a very high cardinatlity maybe 1000. Thats the new addition to our page that results in this if I remove this filter i get very fast results.

            I did not think we could do anything to the data engine in terms of index, how can i tell Hyper to index that table or optimize it further

            • 3. Re: Query performance
              Jeff Strauss

              1000 unique values while a lot, doesn't seem overly excessive unless the length of the value is really long.  What if you change the quick filter type to a text box instead of drop down, does this help at all?  Also, Hyper should automatically index / optimize so there's not much to be done here.  Are you able to attach your workbook by chance?  One other thought is to try and add a context filter on the ones that are quick performing, you can just right click and set this within the filter shelf.  Also, turn off "show only relevant values" for the quick filter setting and see if this helps.