I am facing an issue when trying to add columns from two different data tables into a visual. Appreciate if anyone can help me out on this.

[Shehryar Khan]

I am facing an issue when trying to add columns from two different data tables into a visual. My data source is Microsoft SQL Server and I have imported one fact table and one dimension table into my Spotfire file. I have a column match between these two table, and have created a relationship between the two tables as well based on the same field used in the column match (attached a screenshot of both). However, Iam unable to use any otherfield from my dimension table that was not used inthe column matchwith my fact table when creating a visual.The error I am getting is as follows:

An expression used on an axis in the visualization is not valid with the current setup. Consider selecting a column from the main data table or specifying an aggregation method on that axis (I have attached a screenshot of the error as well)

[David Boot-Olazabal]

Hi Shehryar,

I think you can only use the matched column in the visualization when adding columns from multiple data tables into a visualization.

I have tested this in Spotfire 11.8, and I can add in matched/not matched state columns from multiple sources. But in the unrelated case, I will get a warning and a flat line for one of the columns. While in related state, there is no warning and both columns display their values properly.

When you add a relationship between the two data tables, this makes no difference.

Kind regards,

David Boot-Olazabal

[Anders Gavare]

The key phrase in the error message is "Consider selecting a column from the main data table".

Try changing the Data table in the legend (the main data table) to V_MIG_DimProduct. Then you can pull in

Sum([sales Query].[sales]), Sum([sales Query].[units])as measures from secondary tables on the Y axis. However, depending on which other columns are used in the plot (such as V_MIG_DimProduct.Channel), data may be "smeared" if there is no column match for that column, which may or may not be what you want. (It depends on how the data is structured.)