build time issue while using Elastic cube as data source in other cube

Answered

Comments

3 comments

  • Avatar
    Michael Becker

    Hi Ayush,

    my initial thinking would be that if you have both of those ElastiCubes on the same machine, they utilize the same resources (memory, CPU). In first scenario database has its' own resources (memory,CPU) on different machine. I'm just suspecting that this might be the reason for the difference. Other thing you might consider is changing chunk size for the import. By default it is set to 100k, but you can go so settings in Sisense Server Console (top right corner there) and in those settings(expand for all settings) you can change chunk size. Maybe changing that to lower/higher value will improve performance.

    Regards,

    Michał

    0
    Comment actions Permalink
  • Avatar
    Sam Ovenshine

    Hi Ayush,

    I also want to point out that the Sisense's ElastiCube connector is designed to port small tables over from one ElastiCube to another. It is not designed to handle large data sets or exist as a full-fledged connector. So as counterintuitive as it sounds, the behavior you are experiencing is in line with its design, and for a table of that size, I would recommend using your original connector.

    Thanks,

    Sam

    -1
    Comment actions Permalink
  • Avatar
    Manish Kumar

    I would agree with Ayush.  reading from an ElasticCube should be the fastest method to bring data into another cube.  This is true with QLiksense, i.e QVDs.  The same should be true with Elastic Cubes.  There are many reasons to want this: Sharing tables that are used by many cubes, manage partial loads versus full loads etc.  Sisense really needs to up there game ETL functionality.  

    2
    Comment actions Permalink

Please sign in to leave a comment.