how do you manage new columns in a tabular table

  • Hi we run 2019 ssas std.  Yesterday i imported my tabular project to vs 2022 and changed data source and target to be our dev box.

    One of the tables , "sales" has code behind it that looks like this "select * from vfactsales".   I'm just the messenger here.

    moments before i imported, i changed vfactsales to include a handful of new columns.  I didnt see those columns in the design area nor the area along the right with all the metadata where you can see tables , columns etc.

    i poked around a little to see how one can add the new columns to the cube.   I stumbled on a red exclamation point in an editor type window where one might develop queries.   Not sure but i think running it caused my columns to appear.

    so what is the appropriate way to add new columns to a table in tabular?  it doesnt seem intuitive or "automatic".   And had i changed vfactsales in production to include the new columns, before managing the tabular db, might i have been (un)pleasantly surprised that my cube got built automatically because of the select * with the new columns?   I can hear the advice now , list all the columns in that select, get rid of the *.   currently, because of space issues we had way back,  we process this cube in pieces , i think dims separate from fact in separate steps but it is the equivalent of a full every day.

     

  • Although it does not directly address your question, my suggestion would be that you learn how to use Tabular Editor (link) rather than VS to modify your tabular models. It is a much better experience.


  • This was removed by the editor as SPAM

  • thx phil, as always very useful info.

  • thx morgan3445.  there is no edit dax button in the properties pane of my sales table.

  • stan wrote:

    thx morgan3445.  there is no edit dax button in the properties pane of my sales table.

    The Morgan3445 post contained a spammy link, so I reported it.


  • thx phil, i saw that today and wondered.

    for the community i think i was being stupid.   vfactsales didnt have the handful of new columns in it.   i only noticed when i got an error clearing and processing.   i created vfactsales1 with the new columns.   rather than waiting to see if my vs project then added the new fields immediately to the list of columns (it tries to reload locally after such a change),  i used the cube definition (i believed these are called bim)  and recreated the cube db after replacing every occurrence of vfactsales in that bim script with vfactsales1 and executing the script in an xmla pane .    things look pretty good.   i've asked my user to report any performance hits.

Viewing 7 posts - 1 through 6 (of 6 total)

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