Why can't I ad a Distinct Count measure in an OLAP cube without a non-Distinct one?

Discussion in 'SAS (Statistical Analysis Software)' started by Sander Burggraaff, Sep 18, 2006.

  1. What's the reason for having to select a non-distinct count when you
    want to add a distinct count in an OLAP cube?

    I have a SAS dataset that is an extract from the metadata server log.
    It contains observations for each new client connection made by an user
    including an user id and a date & time stamp.
    I want to know the distinct users per day or month or whatever and I
    want to be able to view each client connection with it's timestamp in
    the detail view of the cube I created.

    The ideal situation would be using that dataset as input for the cube
    and then adding a distinct count on the 'user id'. But I can't
    because OLAP Cube Studio won't let me add such a count unless I also
    select a non-distinct count. But I don't have any other measures to
    select.

    A work around is adding a variable 'number of client connections'
    to the SAS dataset which always holds a value of 1 and select that as a
    summarizing measure. Then I can add the distinct count. But it seems a
    bit silly that I have to create something because I want something
    else.

    Why? What makes a distinct count so special that there has to be a
    non-distinct count measure before I can add a non-distinct one?

    With regards,


    Sander Burggraaff
     
    Sander Burggraaff, Sep 18, 2006
    #1
    1. Advertisements

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.