Dimension Table used in Multiple Cubes, Msut I change all related SSIS Pkgs and Cubes?

  • Hello,

    System: SQL Server 2014, SSIS, SSAS, VS2013, SSDT, and Attuity 3.0 Teradata, and Tabluar Module Partions.

    I have one certain Dimension table being used in several different cubes wired into their diagrams. If I Alter in a new column to that certain Dimension table, update the SSIS package that works with that Dimension table for column Alteration, and then update the Cube design with that new Dimension:

    * Can I alter that one certain table, update that one cubes SSIS package Cube, and all the other cubes using that same one Dimension table will keep running without error?

    * Or will I need to also update all the other SSIS packages and Cube designs that also use that one updated Dimension table or they will error on me.

    Thanks, JPQ

  • As far as I am aware, adding a column to a back-end dimension table will not break your cube build. The next time you update metadata, the column will be added to the model if you are connecting to tables directly.

    If you'd like to avoid any potential impact from schema changes, I'd recommend using queries to extract the data from your source tables.

  • Martin Schoombee (11/4/2016)


    As far as I am aware, adding a column to a back-end dimension table will not break your cube build. The next time you update metadata, the column will be added to the model if you are connecting to tables directly.

    If you'd like to avoid any potential impact from schema changes, I'd recommend using queries to extract the data from your source tables.

    What I'm finding, working with cubes that interact with the Tabular Module Partition tables, and are reading SSQL views with "SELECT Col1, Col2, Col3, etc..." and not "SELECT *". and uniqure views created for this cubes use, based off a tables where it's data is used in other cubes, but with their own custom named view...., this I think saves me.

    So I made a change yesterday to the base SSQL table to add in the new column, as it was already in the DW table that feeds it, and then added the new column to the custom view in the target cube. And let it run the daily jobs the next morning. I'm waiting for runs to complete now.

    I presume all will be fine, then I will go in and modify the cubes partition and add the new dimension column, check the table in diagram and make sure all is wired up.

  • Can I undo an old version, when I don't have version control software? Is there a way to 'reset', or 'clear' a cube of it's partition memory if it were?

    I'm finding that the partition tables and sprocs used to create, process and delete partitions, and what previous months they have processes, is remembered in the cube.

    It is that memory of partitions processed I'd like to clear from the cube in production, I'd like to keep everything else, and then run it, and it would process in Jan16 - Nov16 data from the source.

    If I can't, then does this mean that I have to reconstruct (copy/import) a whole new cube identical to what I have (and the SSIS packages related too), but different name, insert the necessary values into the partition tables, and then run it run like as a brand new cube?

  • quinn.jay (11/10/2016)


    Can I undo an old version, when I don't have version control software? Is there a way to 'reset', or 'clear' a cube of it's partition memory if it were?

    I'm finding that the partition tables and sprocs used to create, process and delete partitions, and what previous months they have processes, is remembered in the cube.

    It is that memory of partitions processed I'd like to clear from the cube in production, I'd like to keep everything else, and then run it, and it would process in Jan16 - Nov16 data from the source.

    If I can't, then does this mean that I have to reconstruct (copy/import) a whole new cube identical to what I have (and the SSIS packages related too), but different name, insert the necessary values into the partition tables, and then run it run like as a brand new cube?

    Answer, go to the SSAS, find the cube, go to partitions, and go to the table that has the partitions created, and delete all or the ones in question. Then delete some or all the records related to the cube in question from ssaspartition, ssaspartitionmanager. Now run the package again, and it should create fresh all the partitions.

Viewing 5 posts - 1 through 4 (of 4 total)

You must be logged in to reply to this topic. Login to reply