1 Reply Latest reply: Apr 4, 2015 3:36 AM by wrvs fsbv RSS

    Qlik Sense 1.1 - long connect times with Direct Query

    Mike Woodward

      Hi,

       

      I'm experiencing extremely long connect times using Direct Query and I wonder if anyone has any suggestions for improving performance.

       

      I'm using Qlik Sense 1.1 connecting to an Amazon RDS database. Because the table I want to read is so big (100 million plus rows), I'm using Direct Query (in-memory won't work for me). It's taking 12 minutes for my load script to get to the point where it's reading in fields. There's a big delay before each field is loading too - 4 to 5 minutes before the field is loaded. On the same machine, I'm using SQL directly and I'm getting data back in a fraction of that time.

       

      Here's my load script.

       

      LIB CONNECT TO 'RDS';

       

      Trace Name of table;

       

      [Name of table]:

      DIRECT QUERY

      dimension

         key_field_id,

          `field_id2`,

          `field_id3`,

          value

      FROM `database`.`table`;

       

       

      Anyone got any ideas for why it's taking 12 minutes before I'm seeing data being read in? Any idea why there's a delay of 4 to 5 minutes before each field is read in?

        • Re: Qlik Sense 1.1 - long connect times with Direct Query
          wrvs fsbv

          You can try connection pooling. When a connection to a database is made a new connection is made and destroyed every time a user connects and disconnects.

          Database service provider can implement a pooling technique that can allow  multiple connection objects from a resource pool to be shared transparently among the requesting clients. Basically, connections are not destroyed but are released after use into a pool of connections.

           

          The load script syntax to setup the pooling capability is as follows:

          SET DirectConnectionMax=10;

          The default setting is 1.