Table column shows as blank until I reorder the columns and then values appear

[Anonymous]

I have two related tables:

• The support_group column belongs to TABLE A
• The OS Platform and Operating System columns both belong to TABLE B.

The following takes place in the Report view...

If I want support_group to appear after any of the fields in TABLE B it will contain blanks even though it really does have data in memory (I know this because I have a filter that selects records based on a certain string in that column). Note also that if I try exporting this table, the .CSV file also propagates the error and the support_group column contains blanks:

However, if I move support_group (in TABLE A) to the left such that it appears before all of the columns in TABLE B, then the data in the column suddenly appears (data is masked, but you can see something is in there, and if I now perform an export, the column is populated in the .CSV file):

This seems to be a defect; I am unable to find any reference to this limitation.

[Anonymous]

@AllisonKennedy 

I found a reference to the order behavior embedded in an article on a different topic; it's too bad they don't make this easier to be aware of as a fundamental note in the Visualization Pane: 

https://docs.microsoft.com/en-us/power-bi/create-reports/desktop-show-items-no-data

[Anonymous]

@AllisonKennedy The tables were joined as M:M; today I created a simple report with a small 1:M-based dataset and the drag-and-drop reordering within the Visualization pane's Values bucket worked as I expected, with no blanking of data when I executed the same column reordering steps.

It seems odd that the Visualization pane's Values bucket would have an effect on the querying of the underlying data model; i.e., I thought it was simply a distinct presentation layer that logically reordered the Table visualization's output because clearly the Filter pane's configuration was using that underlying model's query data to select on the records. This means the filters could "see" the actual data in that column even though the Table visualization presented it as blank when it was in the "wrong" position. Is there any documentation available that warns of using field reordering in the Visualization pane with M:M tables? I could not find any after extensive searches.

[AllisonKennedy]

@Anonymous You're right, it's difficult to find documentation on this 'feature'. I'll add it to my list of blog ideas, but please post any links for reference if you do find any, and I'll try to do the same.

[AllisonKennedy]

@Anonymous  this is to do with how the tables are joined - you're creating the join in the table visualization and if you do a right outer or left outer it changes based on the order of your fields. How is table a related to table b?