1 Reply Latest reply on Mar 14, 2018 10:24 AM by Deepak Rai

    Data Padding (densification) for showing labels for missing values

    Alex Bar



      I need to show all labels in a crosstab even when there are missing values in the underlying db.


      I have created a Custom SQL Query that duplicates all rows of a table in a db and creates observations for (almost) each distinct value of the dimensions in the view as per Missing values in dimensions


      However, I am missing one additional dimension, which comes from another table of the db (that is in left join with the padded table in the datasource).


      How can I further perform data padding for that specific dimension, so that I have one observation for each possible combination of all dimensions in the view, including the one on the filter shelf coming from another table? Right now the cross tab does not show labels for rows when I filter the view by such dimension because of the lack of relevant observations .


      I cannot share the workbook because of corporate requirements.